Description:
There is a new function, meiPlatformMapController(), to set the PCI window address while using 2 ZMPs that are referring to each other's memory.

Description

Support was added for resetting the Mitutoyo encoder on a S200 drive.

Description
The MPI now supports MEIEventTypeMOTION_START to complement MPIEventTypeMOTION_DONE.

Description
Convex CSDM4-SQ drive now supports firmware downloads and velocity mode. These features are supported in Convex CSDM4-SQ firmware version 1.10.

Description
The following changes were made to the LAMV2ParamRangeCheck() function to support LAM0 V2 drive firmware:

The range check values for the MICONT parameter were changed from 5-1750 to 50-17500.

The range check values for the MIPEAK parameter were changed from 5-3500 to 50-35000.

The range check values for the MFBDIR parameter were changed from 0-7 to 1-7.

Additionally for the MV2ParamRangeCheck() function, range checking was added to the following parameters:

KVAFR
VF0, VF1, VF2, VF3, VF4, VF5, VF6
CONFIG
I2TSAMPL
ML0, ML1, ML2, ML3, ML4, ML5, ML6, ML7

Description
S200 drive module now reports the AuxFBMitFlt register and the AuxFBMitAlrm register values, if the device feedback is set to the Mitutoyo interface when reporting the drive faults.

Additionally, Mitutoyo feedback fault codes were added as an extension of existing S200 drive fault codes. Previous S200 fault codes were limited from 1 to 19 and were extended to 27.

Note: Fault codes 20 to 27 are dedicated to Mitutoyo feedback.

Description
Convex CSDS1 drive was limited to Torque Demand mode support. Velocity Demand mode support for Convex CSDS1 was added.

Note: Use drive firmware version 1.10 for Velocity Demand support.

Description
Support for Mitutoyo encoder feedback type was added to the MPI.

Note: The S200 drive firmware version must be greater than or equal to 3.2l and the FPGA version must be greater than or equal to 0x0400, branch 0x020E. The Mitutoyo feedback type enumeration types are used for drive parameters get\set. Otherwise the old enumerations are used for drive parameter get\set.

Description
Support for Convex CSDS1 - SQ was added to the drive module library.

Description
Support for S200 drives S22460 and S24860 (AC5 and AC6).

Description
Added support for Convex CSDM4-SQ drive.

Description
Added support to MPI for the following new GMS modules:
GMS STC-SQA2
GMS STC-SQA2-Rev2
GMS STC-SQA2S
GMS STC-SQP2
GMS STC-SQP2C

How do I use this feature?
Contact GMS for more information.

Description
Added Yaskawa SGDS drive parameter support for the following parameters:

• Pn50A
• Pn50B
• Pn50F
• Pn6F0

These parameters are available with meiSqNodeDriveConfigGet/Set.

How do I use this feature?
The support situation of a new parameter depends on the firmware version of the drive. If the drive does not support the parameter, the data value will be -1.

Description
Velocity mode support is added for Yaskawa SGDS drive. Parameters pn406, pn50B, pn50F, pn581, pn6F0 and extended Parameters 1008-Motor rated speed 1009-Motor max speed 100A-Motor Type String support has been added.

How do I use this feature?
For more information, please contact Yaskwawa.

Description:
CacheService object creates and handles a thread that helps update the Controller’s status cache. The CacheService object is a convenient way to have a program periodically update the Controller’s status cache, allowing the MPI methods to avoid the occasional penalty of explicitly updating the cache.

How do I use this feature?
The CacheService object is not part of the standard MPI. In order to use the CacheService Object, the apputil.h header file needs to be included by your code and the apputil library linked to your application.

For more information on this new feature, see CacheService Object.

Description:
In order to reduce the function execution time when polling MPI status attributes in a client/server setting, some status structures will be cached on the client control object. With caching enabled, certain MPI status methods will get status values that are stored locally in a cache. This feature only works for client type controllers.

How do I use this feature?
Caching can be enabled by calling mpiControlStatusCacheIntervalSet(...) to set a cache refresh interval. Once caching is enabled, the supported status functions will automatically refresh the cache if it is out of date with respect to the refresh interval.

For more information on this feature, see Status Caching.

Description:
Support was added to the MPI for the new controller: ZMP-SYNQNET-CPCI-3U-uD-XIO. The control I/O tables were extended to support the meiControlInfo(...) method with the new ZMP controller.

Description:
The firmware transfers internal controller variables to fields of the SynqNet Demand packet each sample. Currently the internal variables are limited to 32-bits (float and long). By adding support for the transfer of the 64-bit type double, the outputs from MechaWare blocks can be transferred to the Demand packet. This feature would not normally be used directly by a customer. It is intended for use by future demand output MechaWare blocks.

A new enumerated type MEIXmpDACInputTypeDOUBLE, was added to the MEIXmpDACInputType enumeration. Code was added to motor.c to support the transfer of this type.

How do I use this feature?
The appropriate DemandChannel[ ].Input pointer is set by the MPI to the address of the controller variable. The value of DemandChannel[ ].InputType is set to MEIXmpDACInputTypeDOUBLE.

Description:
The S200 drive firmware can be now downloaded via SynqNet. The drive Firmware versions 3.0A to 127.9Z and 255.0A to 255.9Z support firmware download via SynqNet. The following versions do not support drive firmware download: 2.0a, 2.1a, 2.1b, 129.0a, 130.0a

Description:
Demand mode can now be switched while the amplifier is enabled, which was not possible before. Only certain mode switches are supported, such as between velocity and torque mode. Motors can only be switched to a mode that the motor supports.

How do I use this feature?
meiMotorDemandModeSet(...) is called with a motor and the demand mode to switch to. Demand mode can still be changed using mpiMotorConfigSet(...), but only while the amp is not enabled.

Description:
Support for the feed-forward terms has been added for velocity mode. In SynqNet velocity mode, the controller will send both a velocity demand and torque demand to the drive. In previous versions, the torque demand was always zero and the velocity demand would be set to the output of the closed-loop control algorithm.

How do I use this feature?
In version 03.04.08 and later, if a motor is in velocity mode the Kaff and Kff terms will apply to the torque demand field, the other terms will apply to the velocity demand field. Please see the PID and PIV block diagrams.

The torque mode feed-forward terms operation has not been changed.

Description:
For Velocity and Position Mode drives that support gain tables, the gain table index will be sent to the drive via SynqNet. This feature allows users to switch gain tables in the controller and the drive simultaneously. The gain index is sent to all drives, even if the drives do not support it. Drives that do not support the gain index will ignore the data. Please consult your drive manufacturer's documentation for details about drive parameters for gain tables.

How do I use this feature?
The controller firmware automatically copies the controller's gain index value to the drive via cyclic data in the SynqNet packets. See Gain Tables.

Description:
New Yaskawa drive-specific methods were added to read the encoder counts per revolution for feedback devices that support this feature. These drive-specific methods will be replaced in a future release with a general purpose interface to read the feedback counts per revolution.

How do I use this feature?
The following methods can now be used to read the encoder counts per revolution for the Yaskawa drives/motors. If a drive/motor does not support the encoder counts per revolution, it will return a value of zero for the counts per revolution.

   sgdsEncoderCountsPerRev(MPIControl  control,
                           long        nodeNumber,
                           long        *countsPerRev);

   sgdsBsEncoderCountsPerRev(MPIControl control,
                             long       nodeNumber,
                             long       driveIndex,
                             long       *countsPerRev);

   sgdzMdEncoderCountsPerRev(MPIControl control,
                             long       nodeNumber,
                             long       driveIndex,
                             long       *countsPerRev);

    control - a handle to a control object.
    nodeNumber - index to the SqNode.
    driveIndex - index to the drive interface, relative to the SqNode.
    *countsPerRev - a pointer to the number of counts per motor revolution.
                              A value of zero indicates the counts per revolution are unknown.

Description:
The new AMC-DPQ series drives are now supported. The following drives are now supported.
   - DPQNNIE
   - DPQNNIR
   - DPQNNIA
   - DPQNNIS

How do I use this feature?
Contact Advanced Motion Controls for more information. (www.a-m-c.com)

Description:
The Kollmorgen SqDC drive will generate a Axis Current Limit Fault when the axis current exceeds the programmed current limit value.

How do I use this feature?
Support has been added to the MPI for the Axis Current Fault.

Description:
The mpiMotorPositionFeedbackOffsetSet(...) and mpiMotorPositionFeedbackOffsetGet(...) methods were added to the MPI to allow user applications the ability to store and read back an absolute position offset value. This value will be stored in the EEPROM on the motor itself. This feature will be supported through the drive interface on selected drives only. These methods can directly access storage on the EEPROM, allowing a position offset to be stored from session to session. In conjunction with mpiAxisOriginSet(...), this feature replicates the ability of certain drives to reset the zero position on absolute encoders.

How do I use this feature?
This feature can only be used with a Kollmorgan S200 drive and an EnDat 2.1-capable motor. The S200 must have AuxFB set to "EnDat 2.1" in order for this feature to work correctly.

Description:
Previous MPI versions would not support local controller I/O for ZMP models T115-0001 and T127-0001. Support has now been added to the MPI for the T115-0001 and T127-0001.

How do I use this feature?
The MPI uses the model numbers to decode the local controller digital I/O bit availability. The local controller I/O can be identified via meiControlInfo(...) and accessed with mpiControlDigitalIn(...), mpiControlDigitalOutGet(...), or mpiControlDigitalOutSet(...).

Description:
Support has been added to the MPI for the following Yaskawa SGDZ-BS Series Drives: SGDZ-BS61, SGDZ-BS62, SGDZ-BS63.

How do I use this feature?
For more information, please contact Yaskwawa (www.yaskawa.com).

Description:
The Kollmorgen CD drive with DSP rev G has a new drive firmware download sequence. To support the DSP rev G, the drive firmware download method was modified and a new kollmorgen_ember.a00 file was updated. The support for the drive firmware download feature on the DSP rev G is not backwards compatible with older MPI versions.

How do I use this feature?
For more information, please see the Kollmorgen CD drive.

Description:
The Kollmorgen SqDC and SqStep drives now support warning codes/messages. Support has also been added for the SqDC2 model.

How do I use this feature?
For more information, please see the Kollmorgen SqDC and SqStep drives.

Description:
The Sanyo Denki R-Series is now supported. The Q-Series support has been expanded to include service channel, drive faults/warnings, drive parameters, etc. In previous MPI versions, the Q-Series support was minimal. If you have a Q-Series drive and upgrade the MPI, please contact Sanyo Denki (www.sanyodenki.com) for the latest drive firmware.

How do I use this feature?
For more information, please contact Sanyo Denki (www.sanyodenki.com).
See also, Sanyo Denki Q-Series and R-Series drives.

Description:
New functions were added to get the lower 32 bits of actual and command position data.

mpiAxisActualPositionGet32(...)
mpiAxisCommandPositionGet32(...)

If only the lower 32 bits of position data need to be read, it's faster to use these new functions instead of using the standard -Get position functions.

How do I use this feature?
For descriptions and sample code, see the links to the function pages above.

Description:
The following drive update rate options are now available on the Glentek Omega Drive. EEPROM option numbers 4-7 were added.

How do I use this feature?
See the Glentek Omega Drive. To determine the update rate or period for the drive, use MEISynqNetTiming.node[].updateFreq or MEISynqNetTiming.node[].updatePeriod.

Description:
Support was added to the MPI for the Kollmorgen SqDC drive.

How do I use this feature?
See Kollmorgen SqDC Drive for more details.

Description:
In a ring topology, SynqNet allows the remaining nodes to operate if one or more nodes fail. Previously, the only way to restore normal operation for all nodes was to reinitialize the entire network, which would disrupt any operating node. In cases of modular machines, it may be necessary to service one or more nodes while the other nodes are actively controling motion and I/O. The new SynqNet HotReplace feature allows one or more nodes to be shut down, serviced, and then brought back to normal operation without affecting the operation of other nodes.

How do I use this feature?
See SynqNet HotReplace for more details.

Description:
Support was added for the Pn600 "Regenerative Resistor Capacity" drive parameter for the SGDS-72/75 drives.

How do I use this feature?
To access the Pn600 drive parameter, use the methods meiSqNodeDriveParamGet(...) and meiSqNodeDriveParamSet(...), or meiSqNodeDriveParamListGet(...) and meiSqNodeDriveParamListSet(...). See the Yaskawa SGDS Drive.

Description:
New methods were added to read and write the axis origin to/from the controller's flash memory. This feature is useful for applications that use ABS encoders. However, you cannot save the ABS encoder offset to the drive's flash memory.

How do I use this feature?
To write the origin to the controller's flash memory, use mpiAxisFlashOriginSet(...).
To read the origin from the controller's flash memory, use mpiAxisFlashOriginGet(...).
NOTE: These methods do not access the origin value of the axis in the controller's dynamic memory.

Description:
The FPGA that is used with SQID nodes, C0FE0030_xxxx.sff, has been extended to detect errors communicating with any of the I/O modules (e.g. DIN32DOU32). The MPI has been extended to report these new error conditions.

How do I use this feature?
See the I/O Faults section. See MEISqNodeStatusIoFaults and MEISqNodeStatus.

Description:
Configurable Demand Modes have been added to the MPI. The new feature should be used when it is appropriate for a drive to change its demand mode. Currently, only the S300, S600, and S1800 drives support multiple demand modes. Check with the drive manufacturer for possible settings. Nodes that do not support a drive interface (ex: RMB-SynqNet) should use ANALOG or use ANALOG_DUAL_DAC for nodes with two DAC's per axis.

How do I use this feature?
See MEIMotorDemandMode and MEIMotorConfig. A drive will default to an appropriate demand mode. Use mpiMotorConfigGet/Set(...) to change the demand mode. See SynqNet Demand Modes.

Description:
The controller's command and actual position registers have been expanded to 64 bit. The MPI uses doubles for command and actual positions, which are limited to 1.7E+/-308 (at least 15 digits precision).

How do I use this feature?
See the "Position Resolution Extended" section of the Important Things to Know page.

Description:
Support was added to the MPI for the Kollmorgen S1800 drive.

How do I use this feature?
See Kollmorgen S1800 Drive for more details.

Description:
The maximum number of controllers supported using a particular OS is now available from a macro, MEIPlatformControlCountMax, in platform.h.

How do I use this feature?
This macro should only be used to determine the maximum number of controllers supported on a single system. See MEIPlatformControlCountMax.

Description:
Support was added to the MPI for the Yaskawa SGDS75A drive.

How do I use this feature?
See Yaskawa SGDS Drive for more details.

Description:
It is now possible to determine whether a non-idle axis and motion supervisor state was caused by an application on the host computer or by a controller action. It is useful to know whether a non-IDLE state was caused by a controller event or a user action. If the action source was the controller, the controller may be queried to find out what event caused the error state. In a multi-threaded or multi-process environment, it can also be useful to know if a another thread or process commanded an action that placed the motion supervisor or axis into a non-idle state.

How do I use this feature?
The action source can be obtained by calling mpiMotionStatus(...) or mpiAxisStatus(...) and then reading the MPIStatus.actionSource value. Use the enumeration, MPIActionSource, to decode the value.

Description:
The XmpBufferData.VolatileBuffer was added to the controller's external memory. This volatile memory buffer is initialized to zero after a controller reset or power-up cycle.

How do I use this feature?
The volatile memory buffer can be useful for setting application flags, which can denote that certain operations have taken place. For example, once an axis has been successfully homed, a flag can be set in the volatile buffer. However, after a controller reset, the flag will be reset to zero, indicating that the axis is no longer homed.

Description:
The following functions have been added to sqNode.h to assist the setup of recorders and user limits with I/O on SynqNet nodes.

sqNodeDigitalInPtr(...)
sqNodeDigitalOutPtr(...)
sqNodeAnalogInPtr(...)
sqNodeAnalogOutPtr(...)

How do I use this feature?
For descriptions and sample code, see the links to the function pages above.

Description:
The meiSqNodeDriveMonitorInfo(...) method was added to the MPI for retrieving drive monitor information. This method retrieves the number of supported drive monitors, the firmware locations of the monitors, a mask and shift value for retrieving the monitors, and a list of possible monitor configurations.

How do I use this feature?
This feature should be used when configuring drive monitors. The sqDriveMonitor utility contains an example of how to read and display the monitor information.

Description:
The compatibility of a client object to a server is now verified by meiClientValidate(...). Compatibility is based on MPI_INTERFACE_VERSION and the equality of members of the MEIRemoteMethod enumeration.

How do I use this feature?
The compatibility check feature is automatically invoked when the client object is validated.

Description:
Support was added to the MPI for the Kollmorgen S600 drive.

How do I use this feature?
See Kollmorgen S600 Drive for more details.

Description:
Support was added to the MPI for the Kollmorgen S300 drive.

How do I use this feature?
See Kollmorgen S300 Drive for more details.

Description:
A new function has been added to the MPI which can be used to determine the size of the controller's frame buffer and how many frames are currently in the frame buffer for a particular axis.

How do I use this feature?
See meiAxisFrameBufferStatus(...) for sample code and implementation details.

Description:
The controller firmware performance can be measured with meiControlStatistics(...). This method will read the maximum and average times for the foreground and background controller firmware tasks. The controller performance counters can be cleared with meiControlStatisticsReset(...). This feature is useful for determining the controller processor load. It can also help determine the maximum sample rate and number of available resources. For more details, see Sample Rate and SynqNet Controller Performance.

How do I use this feature?
meiControlStatistics(...) | meiControlStatisticsReset(...) | MEIControlStatistics
See the XmpStats1.c sample application.

Description

The number of axes per motion supervisor was expanded from 16 to 32. To support this change, the MEIMotionConfig{...} structure was changed from MEIMotion.axisNumber[16] to MEIMotion.axisNumber[32]:

typedef struct MEIMotionConfig {
char userLabel[MEIObjectLabelCharMAX+1]; /* +1 for NULL terminator */
long axisCount;
long axisNumber[MEIXmpMAX_COORD_AXES];
double blendLimit;
} MEIMotionConfig;

WARNING!! You must re-compile your application if upgrading from a previous MPI version 03.04.xx. If you do not recompile your application and call mpiMotionConfigGet/Set(...), unstable behavior may result.

Note: If 16 axes or more is required per motion supervisor, use mpiMotionConfigGet/Set(...) to configure up to 32 axes per motion supervisor.

No application changes are necessary. You must re-compille your application if upgrading from a previous MPI release.

Description

The Convex CSDS1 drive occasionally fails with a watchdog timer fault because of defective serial data between the SynqNet FPGA and the drive memory.

The problem was caused by a SPI timing violation and was corrected by reducing the serial clock rate from 25 MHz to 12.5 MHz.

Description

In previous releases, when the mpiSgnodeDriveparamStore(..) fuction was called to store the drive parameters into the NVRAM of the LAM drive, the function returned the MEISqNodeMessageRESPONSE_TIMEOUT message. This was caused because the time required to store the parameters into the NVRAM drive was insufficent. This issue was corrected.

Description
If client/server communication is disconnected (broken cable, server.exe exit, etc) the MPI will return a CONNECTION_CLOSED or COMM_ERROR, and try to re-connect. On the next MPI call, if the re-connection is sucessful, the client/server operation will resume.

If the connection failure occurs while the client is reading data, the application can continue with no problems. If the connection failure occurs while the client is writing data to the controller, the data might be partially written. Thus, the client application will need to repeat the previously failed MPI calls.

Description
Added support for new Yaskawa drive parameter Pn581 for the following drives: SGDS-75A, SGDZ-BS51, SGDZ-BS61

Description:
The following functions have been optimized for client/server operation:

• meiSqNodeDriveMonitor
• mpiMotorAmpEnableGet
• mpiMotionStatus
• mpiAxisActualPositionGet
• mpiAxisPositionError

These functions now run remotely on the server. If you are upgrading from a previous release, make sure to upgrade the MPI DLLs on both the client and the server.

Description:
AMC DPQ-Series drives now have three 32-bit words that contain the drive fault and warning codes. To support the additional drive faults/warnings, all three words are read when the meiMotorAmpFault(...) or meiMotorAmpWarning(...) is called.

Description:
WARNING!
Between MPI version 03.04.07 and 03.04.08 the order of the MEIXmpAlgorithm enumeration changed. This change was made for internal performance optimization, but unintentionally affected the MPI interface for the MEIFilterConfig{...} structure.

If your application uses the MEIXmpAlgorithm enum, please be sure to re-compile your application with the 03.04.08 library and header files. If you do not recompile your application, the filter type may not be as expected and a motor may become unstable.

03.04.07

typedef enum { 	
     MEIXmpAlgorithmINVALID = -1, 	
     MEIXmpAlgorithmPID, 	
     MEIXmpAlgorithmPIV, 	
     MEIXmpAlgorithmNONE, 	
     MEIXmpAlgorithmPIV1, 	
     MEIXmpAlgorithmUSER, 

03.04.08

typedef enum { 	
     MEIXmpAlgorithmINVALID = -1, 	
     MEIXmpAlgorithmNONE, 	
     MEIXmpAlgorithmPID, 	
     MEIXmpAlgorithmPIV, 	
     MEIXmpAlgorithmPIV1, 	
     MEIXmpAlgorithmUSER, 

Description:
meiSqNodeInfo(...) for a failed node would return MEISqNodeMessageRESPONSE_TIMEOUT as it attempted to send a service command to read the switchID. Although, a service command timeout is the correct error, this error prevents the user from reading the other fields. The meiSqNodeInfo(...) was modified to only send service commands to the nodes if they have not failed. If the node is in a failed state, the SwitchID is set to an invalid value (-1).

Description:
Parameter range checking was added to mpiMotionConfigSet(...) to protect for
invalid values. mpiMotionConfigSet(...) will return MPIMessagePARAM_INVALID if the decelTime.stop or decelTime.eStop value is negative or greater than 31 bits of controller samples. If zero or less than one sample is specified, a one sample value will be applied.

Description:
The Yaskawa SGDZ-MD Drive motor fault mask was expanded to include MEIMotorFaultMaskAMP_NOT_POWERED by default. This change will cause a dedicated Amp Fault if the Amp Not Powered drive fault bit is active.

Description:
Yaskawa has added new drive parameters to the SGDS drive, which the MPI previously did not support. An updated .dm file was added for support, and the parameter checking code in the drive module was updated to support the new parameters. See the Yaskawa SGDS Drive section for the list of new drive parameters.

Description:
The Yaskawa SGDS motor fault mask was expanded to include MEIMotorFaultMaskAMP_NOT_POWERED by default. This change will cause a dedicated Amp Fault if the Amp Not Powered drive fault bit is active. Changes to application code are not required.

Description:
During SynqNet initialization, there are several steps where the MPI must wait for a time period for the network hardware. This would cause up to 100% CPU usage for several seconds. To reduce the CPU usage, the wait logic was optimized to sleep for the long wait periods and then poll for short intervals. This eliminated the heavy CPU usage without increasing the execution time for mpiControlInit(...) or mpiControlReset(...).

Description:
In previous versions, the drive firmware for the Yaskawa SGDS and SGDZ drives did not allow the multi-turn reset command to be executed while the network was in SYNQ mode. It required the SynqNet network to be shutdown when performing an mtreset. The multi-turn reset function for the Yaskawa SGDS and SGDZ-MD drives has been modified to perform an mtreset without shutting down the whole SynqNet network. Now, when the multi-turn reset is requested, only the node associated with the drive is shutdown (ASYNQ mode) and restarted. Upon successful completion of the multi-turn reset, the node will be brought back to SYNQ mode.

Description:
The SFD device contains useful information about the motor that is connected to the drive. This information is now available as new S200 Drive read-only parameters.

Description:
For path motion (Splines, PVT, PT, etc.), the MPI creates the frames used by the motion and then writes them to the controller. Typically, each point in the path has a corresponding frame for each axis of motion. For path segments with a large number of points (large number of frames), there is an initial group of frames loaded when the motion is started. More frames are loaded as a result of interrupts (events) generated when the axes' frame buffers run low on frames.

In previous versions, the number of frames loaded was always half the capacity of the frame buffer and the "low frame" trigger was hard coded at 32 frames. Although this limit is usually sufficient for the default buffer size of 128 frames, larger buffers could benefit from increasing these limits. In addition, in cases where a backup on path was not needed, the number of initially loaded frames could be larger.

This change will be most useful for customers streaming path (PVT, etc.) points at a high rate. To take advantage of this feature the frame buffers for the axes involved in the streaming motion should have increased frame buffer sizes (larger than the default of 128 frames). Increase the axes' frame buffer sizes with mpiControlConfigSet(...) or the meiConfig utility. Be sure to retain the parameter for path motion as FALSE.

Description:
The AmpNotPowered fault was added into the default MEIMotorFaultConfig.faultMask logic due to changes in the S200 FPGA brake logic. The "Amp Not Powered" fault bit will now trigger the motor's AmpFault when the Brake is applied directly by the drive. See FPGA issue FP523 for details.

Description:
The motor configuration provides a "Brake" source for the general purpose I/O for MEI RMB node types. The Brake logic for the following node type was incorrect: MEI RMB-10V2-SSI with C0FE002C FPGA

For product consistency, the MEI RMB-10V2-SSI was changed to have the same polarity as the standard RMB-10v2 (C0FE002C FPGA).

WARNING: If you are upgrading from a previous MPI software release where you're using an RMB-10V2 with C0FE002C, and are using a Brake source for a general purpose motor I/O, the logic will be opposite in 03.04.00 and later releases.

Description:
In previous versions, the cable length discovery technique was inaccurate for networks with 6 or more nodes. The inaccuracies were caused by an incorrect measurement and calculation. The addition of more nodes to the network caused additional uncertainty in the time measurement, which caused inaccuracies in the cable length calculations. To reduce the inaccuracies, the cable length time measurements now occur between the controller and node or between two nodes. Thus, adding nodes to a network will no longer cause additional inaccuracies to the cable length calculations.

Description:
A potential problem existed with the Control word in frames being or'd into the axis status. Some frames are not meant to have Control words and are supposed to execute in one sample. However, there are strange cases where the user can get the axis to stop on a frame with no Control word and have that frame be active during the background cycle. When this happens, the axis status code OR's in the data in the place where a Control word would be. So, depending on the data at that location, anything could happen. As a precautionary measure, the Control word in the frame to be OR'd with Axis.Status was eliminated.

Description:
A check has been put into place to ensure that Pulse engines are present before configuring a Motor type to Stepper.

Description:
To be consistent with the rest of the MPI the data types for specifying positions for cam moves have been changed from a long to a double.

See MPIMotionCam and MPIMotionAttributes.

Description:
The S200's default secondary feedback was changed from MEIMotorEncoderTypeQUAD_AB to MEIMotorEncoderTypeDRIVE. The default for the S200's Aux Feedback is now Drive Feedback instead of Quadrature.

Description:
The MEIEventStatusInfo structure was expanded to handle 64 bit data. New default data structures were added and defined to overlay the 10 word data buffer that is returned by an event. By default, the new addressing was designed to match the new overlays.

NOTE: If the defaults are changed, the MPI overlays will not be accurate and cannot be used to extract data from the buffer.

See MEIEventStatusInfo.

NEW:
typedef typedef struct MEIEventStatusInfo {
    union {
        MPIHandle  handle;  /* generic */
        MPIAxis    axis;    /* MEIEventTypeAXIS_FIRST ... 
                               MEIEventTypeAXIS_LAST - 1 */
        long       node;    /* MEIEventTypeCAN_FIRST... 
                               MEIEventTypeCAN_LAST - 1  */
        long       number;  /* MPIEventTypeMOTION 
                                    MPIEventTypeMOTOR_FIRST... 
                               MPIEventTypeMOTOR_LAST - 1

                                    MEIEventTypeMOTOR_FIRST ...
                               MEIEventTypeMOTOR_LAST - 1 */
        long       value;   /* MPIEventTypeEXTERNAL */
           } type;

           MEIXmpSignalID signalID;

           /* Contents of addresses specified 
           by MEIEventNotifyData{} */
    union {
        long sampleCounter;
        struct {
           long sampleCounter;
        } motion;
        struct {
           long       sampleCounter;
           long       positionError;
           MEIInt64   actualPosition;
           MEIInt64   commandPosition;
        } axis;
        struct {
           /* Data associated with the CAN event. */
           long data[4];
        } can;
        struct {
           long      sampleCounter;
           long      dedicatedIn;
           MEIInt64  encoderPosition;
        } motor;
        long word[MEIXmpSignalUserData];
     } data;
} MEIEventStatusInfo;

Description:
SynqNet Timing improvements were made to improve node control latency. For 16 kHz drives (the majority of SynqNet nodes), the new scheduling algorithm results in control latency times identical to previous MPI releases. For other drive update rates, and for non-periodic nodes (RMBs), control latency may change.

Since a change in control latency may affect system performance, special attention is recommended for these node types when upgrading from previous MPI releases.

Please see the SynqNet Performance Compatibility section of the online documentation for more information.

The controlLatency sub-structure in MEISynqNetTiming was removed. The calculationTime and calculationSlack were promoted to the MEISynqNetTiming structure.

OLD:
MEISynqNetTiming.controlLatency.calculationTime
MEISynqNetTiming.controlLatency.calculationSlack

NEW:
MEISynqNetTiming.calculationTime
MEISynqNetTiming.calculationSlack

The demandLatency and feedbackLatency were moved to the new per node sub-structure of the MEISynqNetTiming structure.

OLD:
MEISynqNetTiming.controlLatency.demandLatency
MEISynqNetTiming.controlLatency.feedbackLatency

NEW:
MEISynqNetTiming.node[].demandLatency
MEISynqNetTiming.node[].demandLatency

The total latency was promoted and renamed to controlLatency and the latencySlack was promoted and renamed to latencyOverhead.

OLD:
MEISynqNetTiming.controlLatency.total
MEISynqNetTiming.controlLatency.latencySlack

NEW:
MEISynqNetTiming.controlLatency
MEISynqNetTiming.latencyOverhead

Problem/Cause
The problem was caused by new code meant to permit motionModify of path motion.

Code was added to the MPI checks to see if a host frame buffer exists when motionModify is called. If it existed, it was flushed and deleted.

The problem was that the check was incorrect and that the code relied on the user passing in the exact same params structure used in path motion for the motionModify. The code only worked if the motionModify was called to modify the path motion. If the path motion completed and user called startMotion followed by motionModify, the new code would attempt to flush non-existant host frame buffers. While doing this it would execute a switch statement based on the last move's moveType. Since the last move's moveType was S_Curve and the switch statement is looking for path motion types, the switch statement would return an error.

The hold frame only contains 32 bits of position data. Only the lower 32 bits of the position data is used for comparison. The upper 32 bits of position data is not used in the comparison.

Fix/Solution
The following are required for motionModify of a host frame list.

1) A lock needs to be taken for the motion object in the motionExecute function. This prevents the event manager from dumping more frames down to the controller during the motionExecute function. This is already done.

2) FramesLow and FramesEmpty should be active on the controller until the new modified frame list is loaded and ready. The can be done best by modifying the firmware to disable framesLow and framesEmpty when either motionModify or motionStart complete.

3) Modify the firmware to deal with a stopper frame while scanning the frame list for the frame it will modify. If a stopper frame is hit, there is not enough time to complete the modify so an error is returned to the host. Also, since the framesLow and framesEmpty will still be active in this case, the firmware should trigger an ESTOP.

4) Dump some frames from the host frame list to the controller before loading the motion modify on the controller. This is not absolutely necessary, but it does buy time for the motion modify to execute successfully on the controller. More path frames on the controller translates to more time before the stopper frame is hit. Modification will still occur at the same place regardless of how many frames are dumped to the controller.

5) The fix for MPI2412 relied on the user passing in the exact same structure for params without modifying the path motion sections into the motionModify call. The values needed in the structure are the retain and low limit values. To fix this, the retain and lowLimit values are now stored in the motion object when path motion is used. Now, the retain and low limit information stay alive as long as the host frame buffer exists.

6) Delete host frame buffer.

7) Before flushing and deleting frame buffer for motion modify, check to see that valid frame buffer exists.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
When appending a BSpline move that has different time delta from the previous start or appended move, the appended move time is not calculated correctly by the MPI.

The problem is caused by the time array not properly getting sorted when passed down to the controller.

Fix/Solution
The appended move time is now correctly calculated.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
Memory violations occur when mpiMotionStart(...) or mpiMotionModify(...) is called using the MEIMotionTypeFRAME with 2 or more axes.

This problem was caused by a parameter check routine that used the axisCount to calculate the pointCount.

Fix/Solution
The axisCount was not necessary to calculate the pointCount and was removed.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
The firmware was processing the compensator while the MPI was changing it's configuration.

Fix/Solution
compensator.c was modified to zero out the Dimension field in the compensator structure and wait 1 foreground cycle before writing the compensator's configuration to the controller. Once the new configuration is written to the controller, the Dimension field is restored. This prevents the controller from processing the compensator while the MPI changes the configuration. The firmware begins processing the compensator when the Dimension field is non-zero.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
In earlier releases the default velocity mode parameters were not initialized correctly.

Fix/Solution
The default velocity mode parameters were corrected and now initialize properly.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
The FPGA download function fails to download the FPGA to the node if the node has a old boot FPGA and the network is in string mode.

Fix/Solution
The FPGA download function was corrected to allow downloading with a old boot FGPGA node in string mode.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
Previously when the mpiMotorStatus(..) function was called, the Broken wire status of the motor's encoder was not reported correctly.

Fix/Solution
The code was corrected to use the proper bit mask for checking the status of the encoder.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
When a streaming point move (for example: PVT or Spline moves) are modified to S-Curve, Velocity, or Trapezoidal motion types, it is possible for mpiEventMgrService(...) to report an error and to cause an MPIEventMgr object to stop distributing events.

Streaming point moves send events to the host computer when low on frames to request the host to load more frames. If the application changes the motion type to a S-Curve, Trapezoidal, or Velocity move in between the time a low frame event is generated on the controller and the time mpiEventMgrService(...) can handle the event, an error will occur.

Fix/Solution
The MPI now correctly handles this condition. When the motion is modified to an S-Curve, Trapezoidal, or Velocity move, the host frame list is deleted and any subsequent low frames event associated with the currently executing move is safely ignored.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
Appending a Cam profile with mpiMotionStart(...) causes the profile to always start from the present command position, instead of the last position from the previous Cam profile.

This problem was caused by the MPI improperly calculating the initial point for an appended Cam profile.

Fix/Solution
The problem was corrected by applying the last position from the previous Cam profile as the starting point for an appended Cam profile.

Affects to Application Code
These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

WARNING!!
Starting with version 03.04.16, Kollmorgen S300 and S600 drives are initialized to Torque demand mode by default. In previous releases, they were initialized to Velocity mode by default. The default demand mode for other drives have not changed.

If the user set the motor demand mode to Torque, saved to flash, and then reset the controller, the drive library would override the controller and set the demand mode back to Velocity. This prevents the user from selecting the default demand mode.

Fix/Solution
The problem was corrected by removing the S300 and S600 demand mode special initialization from their respective drive modules.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

When the S200ParamI_LMT_PLUS and S200ParamI_LMT_MINUS parameters are read back from the drive, the drivedata is being multiplied by 100/151. This fraction multiplied by an integer number and assigning to an integer is leading to rounding off. This causes the value to be reported as 1 less than the actual value.

The problem was caused by a rounding off error due to assigning floating value to an integer variable.

Fix/Solution
The calculation was corrected.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

The DiPeak value for the S200 drive S21260 (AC4) was 24.

Fix/Solution
The DiPeak values was corrected to 30.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

The Yaskawa SGDZ-MD drive's multi-turn feedback value cannot be reset when SynqNet is in ASYNQ mode.

The problem is caused when the network state is not verified due to a failed service command.

Fix/Solution
The problem was corrected by adding an internal check for the network state. Multi-turn reset will now work properly in ASYNQ mode.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

The Yaskawa SGDZ-BS multi-turn reset fails if the node number is not zero.

The Yaskawa SGDZ-BS drive module's service command flag bits were incorrectly set at 8 to 15 as the node address. This caused the controller to send multiple service commands when the node number was > 0.

Fix/Solution
The service command node index was corrected in the Yaskawa SGDZ-BS drive module.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause

The S200 drive parameters were incorrectly indexed into the drive's memory. meiSqNodeDriveParamGet/Set and meiSqNodeDriveParamListGet/Set fail to read/write the following drive parameters in the drive memory:

KVI_0 to KVI_7
KVP_0 to KVP_7
SFDMEMORY_XXXX
ENC_OUT_PPR
ENC_OUT_OFFSET

The problem was caused by incorrect indices in the MPI's drive parameter memory map table.

Fix/Solution
The drive parameter memory map table was corrected.

Affects to Application Code

These changes were made to the internal MPI libraries and will not affect customer code.

Problem/Cause
The step motors do not move the correct distance and to the target position if the topology was saved followed by changing the sample rate, saving it to flash memory, and resetting the controller. This occurs when meiSynqNetFlashTopologySave(...) saves a stepper motor setting to flash memory that depends on the sample rate. mpiControlFlashConfigSet(...) saves a new sample rate, but does not update the stepper motor settings.

Fix/Solution
If an application attempts to save a new sample rate to flash memory by using a call to mpiControlFlashConfigSet(...) when the SynqNet topology is saved and stepper motors exist on the network, the MPI will return the following error code:

MPIControlMessageSAMPLE_RATE_CHANGE_WITH_STEPPER_MOTORS.

To save the new sample rate to flash memory, the application must clear the SynqNet topology by using a call to meiSynqNetFlashTopologyClear(...), then save the save the sample rate by using mpiControlFlashConfigSet(...), then save the topology again by using meiSynqNetFlashTopologySave(...).

Problem/Cause
In version 03.04.13, the node switchID is 0x000000FF for nodes that do not support an ID switch. In previous releases, the SwitchID value was 0xFFFFFFFF (-1).

Fix/Solution
The unsupported switch ID was changed back to the original value 0xFFFFFFFF (-1).

Problem/Cause
The SynqNet network is incorrectly left in the "SYNQ" state when a drive does not have valid firmware or if the drive intialization fails. This situation can occur after a controller reset, a network initalization, or saving topology to flash. During network initialization, if a drive intialization failed the MPI would return an error but would not change the state from SYNQ to ASYNQ.

Fix/Solution
The MPI library logic was changed to put the network in the ASYNQ state if the drive initialization fails.

Problem/Cause
For the VxWorks platform, the MEI_ASSERT, MEI_TRACE, and MEI_VALIDATE compile options were set for only the debug builds and not for the release build. This disabled various diagnostic functions and macros such as meiAssert, meiValidate, and msgCHECK on the release vxWorks builds. Makefiles for VxWorks were not updated when MEI_ASSERT, etc. were enabled for Debug builds on the other platforms.

Fix/Solution
Updated the makefiles for VxWorks and other non-Windows platforms.

Problem/Cause
If the topology is saved to flash and then the TxTime is changed or if the TxTime is changed and saved to flash, and the controller is reset, it would cause the step pulse output profile to overshoot the commanded trajectory profile. The problem was caused by the StepControl.CommandCountDelay not being saved to flash when the TxTime was saved to flash.

Fix/Solution
It was fixed by saving the StepControl.CommandCountDelay to flash when the TxTime is saved to flash.

Problem/Cause
When running via client/server, if the axes are not mapped, mpiMotionStart(...) returns MPIMotionMessageAXIS_COUNT. A bug was introduced in 03.04.10 client/server optimization that caused the problem.

Fix/Solution
The problem was corrected.

Problem/Cause
When the SSI encoder is configured with mpiMotorConfigSet, it works properly. Later, if mpiMotorConfigGet(...) is called followed by a call to mpiMotorConfigSet, the SSI encoder no longer works.

The problem was caused by a mask used (0x20) to get the bitcount from the configuration.

Fix/Solution
The mask has been corrected.

Problem/Cause
The eventMask argument to mpiSequenceEventNotifyGet(...) was always 0.

The problem was caused by a missing case in an internal status conversion method.

Fix/Solution
The missing logic was added.

Problem/Cause
The drive monitor parameters for the DP and DPQ series drives are different, but the address information returned by meiSqNodeDriveMonitorInfo(...) was the same for both drives.

The AMC drive module did not provide separate drive monitor information for the different AMC drive models.

Fix/Solution
The problem was corrected in the AMC drive module.

Problem/Cause
In MPI release 03.03.05, if a slave is in STOPPED state, it will follow the master. In MPI release 03.04.07, if a slave is in STOPPED state, it does not follow the master.

The issue can be replicated by:

• Gear two axes.
• Command motion on the the slave and call STOP to put the slave in STOPPED state.
• Command motion on the master and check to see if the slave will follow.

The problem was caused by a modification in the order of calculations when the firmware was changed to use 64-bit position data.

Prior to the firmware change, gear position was always added in to the calculation. After the firmware change, gear position is only added if the axis is not in an ERROR or STOPPED state. After clearing the stop, gearing will resume, but will not jump to catch up with where it would have been if the stop did not occur.

Fix/Solution
The problem was corrected by modifying the code to always calculate the gear position regardless of the axis state.

Problem/Cause:
If mpiMotionAction(...) is called with a RESET argument, then the motion type and params in the motion object would be corrupted. This caused subsequent calls to mpiMotionStart(...) to fail with an Argument invalid error. The cause of the problem was determined to be a bug in an internal function.

Fix/Solution:
A bug in meiMotionAxisListsGet(...) was fixed, but it was discovered that mapping more than 16 axes to a motion object is not supported. Please see MPIMotionConfig{...} for details.

Problem/Cause:
Repeated creation and deletion of the controller object (~33,000 with XMP-Series, ~1900 with ZMP-Series) would eventually cause mpiControlCreate(...) to fail with a memory fault.

The problem occurred because meiPlatformDeviceClose(...) was not unmapping all of the device memory originally mapped by meiPlatformDeviceOpen(...).

Fix/Solution:
meiPlatformDeviceClose(...) now unmaps all of the PCI windows rather than just one of them. Repeated creation and deletion of the controller object no longer results in failure.

Problem/Cause:
If the value of actual position exceeds +/- 31 bits before calling mpiAxisCommandPositionSet(...) to zero the position, the command position value would be improperly calculated.

The problem was caused by the MPI changing the command position by greater than +/- 31 bits. The controller calculates a delta position value that exceeds the range of its 32-bit position counter, causing the 64-bit command position value to be calculated improperly.

Fix/Solution:
The problem was corrected by changing the MPI to incrementally update the command position without exceeding +/- 31 bits per sample.

Affects to Application Code:
No application changes are necessary.

Note: If the position error limit is less than the command position setting delta, the position error limit may trigger immediately after setting the command position. If the position error limit is set to ABORT and cmd=actual is configured, then setting the command position will fail because it will be overwritten by the current actual position.

Problem/Cause:
If the controller foreground/background deltas are greater than 10, two different problems can occur with the calculated trajectory velocity term:

1. Final velocity value could be repeated for multiple samples before being zeroed at the end of a move.
2. Velocity value is zero for one sample during a motion profile.

The problem was caused by a race condition in the controller firmware logic where update frames were occasionally not executed if the axis state was IDLE and the internal move_in_progress flag was occasionally in an improper state.

Fix/Solution:
The problem was corrected by changing the logic to fix the race conditions so that the update frame is executed but not advanced to the next frame if the axis state is IDLE (guaranteeing that the final position is updated at the correct sample).

Problem/Cause:
When a data recorder single-shot START trigger is configured from mpiRecorderConfigSet(...) while the conditions to trigger the recorder are true, the trigger may intermittently fail to activate the recorder. The problem was caused by the race condition in mpiRecorderConfigSet(...). The method wrote the trigger configurations to the controller before the recorder configurations.

Fix/Solution:
mpiRecorderConfigSet(...) has been modified to write the recorder configuration before the trigger configuration.

MPI Version: 03.04.09

Problem/Cause:
During mpiControlInit(...) or meiSynqNetInit(...) a Divide By Zero exception would occur intermittently.

A Divide By Zero exception was being thrown by deprecated code that was being executed, but not used by the MPI. During testing, the Visual C++ v6 runtime library did not catch the problem. However, the Visual C++ v7 runtime library was able to detect a floating point division by zero and generate the exception error.

Fix/Solution:
The deprecated code was removed.

Problem/Cause:
PVT move types would fault (position error limit) with a large position discontinuity when the starting position was greater than 32-bits. This error occurred because the initial positions used to generate the frames for path moves (PVT, B-Spline, etc.) were being incorrectly calculated. Only the lower 32 bits of the 64-bit position were being used.

Fix/Solution:
The PVT calculations were revised to correctly calculate the initial position. Multi-point moves are now correct over the entire 64-bit range.

Problem/Cause:
mpiMotionModify(...) with a velocity move type, caused a one sample discontinuity in the velocity profile. The calculated command position was the same value for two successive samples.

The problem was caused by velocity frames being updated with command position values that were one sample old. This caused the motion supervisor to miscalculate the value of the command position when the modify frames were executed.

Fix/Solution:
The problem was corrected by updating the velocity frame's command position with the present command position value.

Problem/Cause:
Calling mpiMotorConfigSet(...) with demand mode = ANALOG_DUAL_DAC, followed by calling mpiMotorConfigGet(...) would return demand mdoe = ANALOG.

The problem occurred because the internal enum values for MEIXmpMotorModeSelectANALOG and MEIXmpMotorModeSelectANALOG_DUAL_DAC were the same.

Fix/Solution:
The problem was corrected by updating the internal demand mode map values for MEIXmpMotorModeSelectANALOG (0xFFFFFFF0) and MEIXmpMotorModeSelectANALOG_DUAL_DAC (0xFFFFF10) to identify the DAC configuration.

Problem/Cause:
If invalid firmware was downloaded to the controller, the mpiControlReset(...) function did not return with an error. The function would lock up because the timer used to check for the timeout depended on the controller sample counter. The problem was caused by a timeout check during the SynqNet network initialization that used the controller's sample counter.

Fix/Solution:
The timeout check function that used the controller's timer during SynqNet network initialization was replaced with a function that checks the host timer.

Problem/Cause:
The command line Flash utility crashed when trying to erase the binary image of a controller via client/server:

flash -server 192.168.1.74 -erase

The crash occurred because the flash utility called meiPlatformFlashSectorErase(...), which does not have a remote method. Since, meiPlatformFlashSectorErase(...) tried to access a local controller directly, it caused a memory exception error.

Fix/Solution:
The problem was corrected by adding a remote method for meiPlatformFlashSectorErase(...). The new remote method is: MEIRemoteMethodFLASH_SECTOR_ERASE.

Warning: When upgrading to 03.04.09, be sure to upgrade both the client and server software to correct this problem.

Problem/Cause:
For ZMP-Series controllers only, meiPlatformMemorySet64(...) incorrectly changed the data of the src pointer. The problem was caused by swapping the data of the src pointer and then writing it to the ZMP. Therefore, the data of the src pointer was changed.

Fix/Solution:
The problem was corrected by copying the data from the src pointer to a temporary variable, swapping it, and then writing it to the ZMP.

Problem/Cause:
If more than one drive was restarted with HotReplace, errors would sometimes occur after the nodes were initialized to SYNQ mode. The errors were drive dependent, but indicated that a drive specific initialization step had failed. In some cases, errors were returned by the HotReplace functions. In other cases, drives were not fully initialized after being restarted, leaving items such as encoders uninitialized. The problem was caused by a logic error that caused only the first drive being restarted to execute the drive specific inialization routines. After the network attained SYNQ mode, the first drive would then be initialized as many times as there were drives that were restarted. As a result, the other drives were left in a partially initialized state.

Fix/Solution:
The logic was corrected in the MPI to run the drive specific inialization routines for each drive that is restarted.

Problem/Cause:
The software position limits would occasisonally falsely trigger when the motor's feedback crossed the zero position from a positive to a negative position value. The problem was caused by the interruption of the position limit comparison with the lower 32-bits and upper 32-bits of feedback position at the same time the 64-bit position crossed from the positive to negative range.

Fix/Solution:
Fixed to firmware. The 64 bit value is now read twice and compared. If the two values are different, processing of the limit is skipped until the next sample.

Problem/Cause:
meiSqNodeStatus(...) would fail with a service command timeout error when accessing a failed Slice node. This error occurred because meiSqNodeStatus(...) was not checking to see whether or not the node was dead before making a service command to the node.

Fix/Solution:
The problem was corrected by having meiSqNodeStatus(...) check if the node is not in a failed state before reading the status.

Problem/Cause:
If a SynqNet network contained a single Yaskawa SGDZ-MD or SGDS drive, the meiMotorMultiturnReset(...) would fail and return a "SynqNet Hot Restart: Not in Synq state" error. The problem was caused by trying to shutdown a single node network, which causes a full network shutdown. The node re-start would then fail, since a full network re-start is required after a full network shutdown..

Fix/Solution:
The problem was corrected by generating a network shutdown to ASYNQ mode, performing the multi-turn reset, and then re-starting the network.

Problem/Cause:
During SynqNet Fault Recovery, the position delta of the bad packet feedback compensation was zero, which caused a discontinuity in the closed-loop servo control output. The problem was caused by a previous change that was added to the encoder processing code which cleared out the delta value so that if the encoder was not enabled, the axis object would not increment its feedback value. As a result, it would get out of synq with the motor object's feedback value. Unfortunately, this change caused an error in the bad packet compensation because it forced the delta value to zero just before bad packet compensation was evaluated.

Fix/Solution:
The problem was corrected by changing the logic in the firmware. The delta value for the encoder is now forced to zero only if the encoder is not processed. This means that bad packet compensation will work as expected and if the encoder is disabled, the delta is forced to zero, preserving the previous change.

Problem/Cause:
If the controller is configured for 32 MS/Axes/Filters/Motors, saved to flash, and then reset, the motor limit configurations would be intialized to zero. This would cause all actions to be initialized to NONE, disabling all limit actions. The controller firmware initialization code clears memory past the last allocated object (i.e. frame buffers, etc...). When the axisCount was set to 32, the firmware cleared past the end of external memory, wrapping around to the beginning of external memory, and overwrote the motor limit configurations.

Fix/Solution:
The problem was corrected by modifying the clear memory routine in the firmware initialization to not clear memory beyond the end of external memory.

Problem/Cause:
The platform lock mechanism did not work for multiple threads running in a single client process when connected to the controller via client server. Locks were only being taken on the server. They needed to be created and taken on the client as well.

Fix/Solution:
The problem was corrected. The platform lock mechanism now works for multiple threads running in a single client process when connected to the controller via client server.

Problem/Cause:
The Yaskawa SGDS drive parameter Pn212 was previously specified as a 16-bit value in the Yaskawa_SGDS.dm file.

Fix/Solution:
Drive parameter Pn212 was changed to a 32-bit value in the MPI and Yaskawa_SGDS.dm file.

Problem/Cause:
The sequence motion commands (seq1.c sample application) did not work. The problem was caused by extending the position range in the controller from 32-bits to 64-bits.

Fix/Solution:
The problem was corrected in the MPI library.

Problem/Cause:
The controller received spurious CRC errors when the idle cable was tested. The problem was a result of the Rincon not properly ignoring the errors.

Fix/Solution:
The MPI now disables the controller RX on the appropriate port before the cable test, thus avoiding the errors from echoed packets.

Problem/Cause:
The following Dedicated In bits are NOT supported by the S200 drive and are now masked from the user:
    MPIMotorDedicatedInINDEX_PRIMARY
    MPIMotorDedicatedInFEEDBACK_FAULT
    MPIMotorDedicatedInDRIVE_STATUS_9
    MPIMotorDedicatedInFEEDBACK_FAULT_PRIMARY

MPIMotorDedicatedInDRIVE_STATUS_10 is now supported. In previous versions it was missing from the drive module library.

Fix/Solution:
Changes were made to kollmorgen_s200.c. See the Drive FPGA Table: Kollmorgen S200 for a complete list of supported Dedicated I/O bits.

Problem/Cause:
Motor limits are configured via mpiMotorConfigSet(...) or mpiMotorEventConfigSet(...). In previous versions, no argument checking was done on the value of OutputPtr when motor limits were set. A NULL value could have resulted in a controller reset or lock-up. A NULL value in OutputPtr would map to 0x0 in controller memory, which is the location of the reset register.

Fix/Solution:
The MPI now checks OutputPtr against NULL and uses the address of MEIXmpDataNull.Output instead. NULL pointers in OutputPtr are now safely ignored.

Problem/Cause:
mpiControlReset(...) via client/server failed after downloading new firmware to an XMP controller whose flash memory had been previously erased. mpiControlReset(...) was reinitializing the Platform object in the "bad firmware case," which caused a new client to be created, thus losing all of the previous client's locks.

Fix/Solution:
mpiControlReset(...) can now be called client/server even when flash has been erased.

Problem/Cause:
In versions 03.04.02 to 03.04.05, if a boot FPGA was found on a SynqNet network, a service channel error was returned. Motion Console was not able to recognize the error and did not automatically prompt the user to download the correct runtime FPGA.

Fix/Solution:
The return values for mpiControlInit(...), mpiControlReset(...), and meiSynqNetInit(...) have been improved:

1. If ALL node FPGAs are runtime FPGAs and compatible with the MPI, the SynqNet network will change to SYNQ mode and MPIMessageOK will be returned.
2. If any node FPGA is a Boot type, the initialization will stop, the SynqNet network will change to ASYNQ mode and MEISynqNetMessageNODE_BOOT_FPGA_VERSION will be returned.
3. If any node FPGA MAC version is not compatible, the initialization will stop, the SynqNet network will change to ASYNQ mode and MEISynqNetMessageNODE_MAC_VERSION will be returned.
4. If any node FPGA is not compatible with the MPI, the initialization will stop, the SynqNet network will change to ASYNQ mode and MEISynqNetMessageNODE_FPGA_VERSION will be returned.

Problem/Cause:
If a Glentek Omega drive had no runtime FPGA image, downloading the runtime image would erase the drive firmware image from flash memory. To correct this problem in previous versions, the drive firmware would need to be re-downloaded. Glentek drives share the same flash part to store both the runtime FPGA image and the drive firmware.

Fix/Solution:
To correct the problem, the download routines now use the node-specific download function to upload the drive firmware, append the runtime FPGA image, and download the full image.

Problem/Cause:
On the S200 drive, motor digital output, "OOUT1" can be used as either a general purpose output or as Brake. In previous versions, the default configuration was output. If it was configured for a Brake and not saved to flash, the Brake would disengage during SynqNet initialization.

Fix/Solution:
To prevent a brake from disengaging during SynqNet initialization, a new GPIO type, "OUTPUT_DISABLED" was added to keep the output in a disabled state during SynqNet initialization. It will now be used as the default configuration state for the S200 GPIO "OOUT1". Any application that uses the S200 GPIO "OOUT1" should configure the I/O explicitly as OUTPUT or as Brake before using this bit. See the Drive FPGA Table: Kollmorgen S200 for more information.

Problem/Cause:
In firmware version 580B3 (and earlier), MEIXmpLimitData.State was set to 0x10 when the limit's condition block evaluated to TRUE and the Duration had been exceeded. In firmware version 580B4 (and greater) the Duration was ignored.

Fix/Solution:
The operation has been returned to the previous logic, where the limit State is not set to 0x10 until the Duration is exceeded.

Problem/Cause:
The Yaskawa SGDS, SGDZ-MD, and SGDZ-BS drive initializations incorrectly configured the motor object for modulo operation based on the counts per revolution. The modulo configurations caused the motor object position and the capture values to be incorrect. A workaround for this problem is to use mpiMotorConfigGet(...) and mpiMotorConfigSet(...) to set the MEIMotorEncoderModulo.value = 0.

Fix/Solution:
The MEIMotorEncoderModulo{...} configuration was removed from the Yaskawa SGDS, SGDZ-MD, and SGDZ-BS drive modules.
WARNING: mpiMotorFeedback(...) is not supported when using a non-zero modulo value. Capture is NOT supported when using a non-zero modulo value.

Problem/Cause:
If mpiMotorEventConfigSet(motor, MEIEventTypeLIMIT_USER0, NULL, &motorEventConfigXmp); is followed by mpiMotorEventConfigSet(motor, MEIEventTypeLIMIT_USER0, &motorEventConfig, NULL); and the second mpiMotorEventConfigSet(...) would automatically set the action to NONE, causing the the User Limit to be disabled. The problem was caused by an optimization to disable unused motor limits when the action was NONE.

Fix/Solution:
This problem was fixed. The optimization to disable User Limits when the action is NONE has been removed.

Problem/Cause:
The MPI would hang during SynqNet initialization if the controller's motorCount was less than the physical number of motors on the network and the physical motor had ABS feedback and the controller was connected to an S200 drive. The problem was caused by initializing an ABS feedback device on an S200 drive with a controller motor object that was not enabled. The MPI code had a while loop that waited for the firmware to update its values. Since the motor was not enabled, the values were not updated and the MPI was stuck in the while loop.

Fix/Solution:
The problem was corrected by adding a timeout to the while loop. If the motor ABS initialization does not have a controller motor object, the MPI will return a TIMEOUT. To avoid the TIMEOUT error, make sure to set the motorCount with mpiControlConfigSet(...) to the number of physical motors.

Problem/Cause:
meiSqNodeDriveParamGet(...) could not read a string parameter from the Yaskawa SGDS and SGDZ-MD drives.

Fix/Solution:
This problem was corrected.

Problem/Cause:
meiSqNodeDownload(...) did not work with the Yaskawa SGDZ-BS drive when downloading drive firmware. The problem was caused by improper transmit and receive buffer configuration.

Fix/Solution:
This problem was corrected.

Problem/Cause:
mpiMotorEventConfigSet(ENCODER_FAULT) would fail with a service command unsupported error code. This problem was caused by an improper service command configuration in the MPI library.

Fix/Solution:
This problem was corrected.

Problem/Cause:
If the encoder ratio was configured with mpiMotorConfigSet(...), there was a possibility of a hang. The problem was caused by non-atomic 64-bit position reads from the controller. If the feedback values are changing across the 32-bit boundary during the position reads, the calculations in the MPI would not work correctly and the MPI would get stuck in a while loop.

Fix/Solution:
The problem was corrected by changing the position reads and calculations so that they do not require 64-bit atomicity. Additionally, a timeout was added to the while loop, so if the controller fails during the calculations, a TIMEOUT error will be returned.

Problem/Cause:
meiSqNodeDriveParamGet(...) was unable to read the following Yaskawa SGDS and SGDZ-BS drive parameters: MotorCapacity, MotorType, SerialNumber, EncoderType, Customize number, and Electric angle. The problem was caused by an incorrect parameter addresses in the MPI library.

Fix/Solution:
The MPI library parameter addressing was corrected.

Problem/Cause:
The MPI would hang during SynqNet initialization if the controller's motorCount was less than the physical number of motors on the network AND if the physical motor had ABS feedback AND was connected to a Kollmorgen S200 drive. The problem was caused by initializing an ABS feedback device on an S200 with a controller motor object that was not enabled. The MPI code has a while loop that waits for the firmware to update values. Since the motor was not enabled, the values were not updated and the MPI was stuck in the while loop.

Fix/Solution:
The problem was corrected by adding a timeout to the while loop. If the motor ABS initialization does not have a controller motor object, the MPI will return a TIMEOUT. To avoid the TIMEOUT error, make sure to set the motorCount with mpiControlConfigSet(...) to the number of physical motors.

Problem/Cause:
Several new SynqNet timing parameters are calculated when the sample rate is changed. These new timing parameters were not being saved in flash memory by mpiControlFlashConfigSet(...) if the current running sample rate matched the new sample rate. Only the value in dynamic memory (not the flash value) of the sample rate was compared to the new sample rate. The timing values were not written to flash becuase the MPI library function, mpiControlFlashConfigSet(...), did not detect a change.

Fix/Solution:
The value of the sample rate stored in flash is now checked and if the new sample rate is different, all the timing parameters are written to flash. When a new sample rate is set (the flash value is changed) the network is re-initialized and all the new parameters are stored in flash.

Problem/Cause:
If the topology was saved and the controller was powered off/on, the SynqNet network initialization would fail. The problem will not occur on ZMP-SynqNet series controllers. The cause of the problem was that the ASYNQ config buffer and the packet config buffer were not restored from flash during reset. As a result, the firmware was not be able to bring the network to SYNQ mode after a hard reset (power down).

In the case of a soft reset, the data in the ASYNQ config and packet buffers would still be there and things would appear to work normally.

Fix/Solution:
The ASYNQ config and packet config buffers are now explicitly restored from flash during boot up.

Problem/Cause:
If Motion Console was open while another application performed a meiSynqNetNodeRestart(...), certain SynqNet nodes (S200, S1800, and Slice I/O) may have failed to restart. This problem was caused by missing node locks during the restart.

Fix/Solution:
To correct this problem, node locks were added to meiSynqNetNodeRestart(...).

Problem/Cause:
The S200 drive sends the absolute encoder data on both the primary and the secondary encoder. However, in previous versions the primary encoder data did not have the absolute multi-turn data. Instead it had relative multi-turn data. Only the Aux encoder would have the correct absolute multi-turn feedback data. There was a previous initialization fix (MPI 2022) in the 03.04.02 release, which switched the S200 secondary encoder pointer to be the primary encoder value and set the encoder counter to 1 in the controller. As a result, the drive module switched the pointer on the firmware memory. Therefore, when the mpiMotorConfigSet(...) function was called, it overwrote the changed pointer to the default stored in the MEIXMPBuffer memory.

Fix/Solution:
Changes have been made so that the drive module changes the default value stored in the MEIXMPBuffer memory. This will ensure that when the mpiMotorConfigSet(...) function is called, the correct Encoder pointer mapping will occur.

Problem/Cause:
Types of motion that needed frame-low events to reload the frame buffers were causing protection faults (access violation errors). The MPI relied on a valid motion handle being returned for frame low events. An incorrect (or missing) initialization of the event data could have caused an invalid handle to be returned by the event. Access to a motion object using this handle would cause a protection fault.

Fix/Solution:
The MPI now writes the correct motion handle every time the motion (needing frame low events) is started.

Problem/Cause:
The drive module converted the float parameter COMM_OFF to a signed 12bit number and stored it in the drive. When reading back the value, the 12bit signed number was treated as a 32bit signed number without correctly extending the sign bit before converting it to a float, which made the negative COMM_OFF angles appear as large +ve values.

Fix/Solution:
The drive module now reads the 12bit signed COMM_OFF angle and correctly extends the sign bit before converting the number to a float.

Problem/Cause:
During the packet configuration portion of the SynqNet network initialization, controller firmware reads the feedback positions/status and writes the torque demands. The controller firmware writes the torque demands at the time of the packet re-configuration. Since the MPI changed the packet format at the same time that the firmware was writing to the packets, it caused the torque demand value to corrupt the packet data. The "bad" packet was detected by the service channel and caused the MPI return code: 0x1b08 (6920) SqNode: service cmd response timeout

Fix/Solution:
In versions 03.04.03 (and later), the network initialization was corrected by preventing the firmware from writing torque demands BEFORE packet configuration and only allowing the firmware to write torque demands AFTER packet configuration. This change protects both the normal network initialization and the packet re-configuration during network re-initialization.

Problem/Cause:
Although the drive parameters could be successfully stored to the EEPROM of the Kollmorgen S1800 drive, a timeout error message would be returned when calling the meiSqNodeDriveParamStore(...) function.

Fix/Solution:
The parameter storage takes about 30 seconds depending on the number of parameters to be stored. This timing requirement is corrected in the Kollmorgen S1800 drive module so that the meiSqNodeDriveParamStore(...) function will wait long enough to store all the parameters.

Problem/Cause:
When using MPIMotionTypePVT with MPIMotionAttrMaskAPPEND and mpiMotionModify(...) to add frames to an already existing set of frames, the BufferLowLimit did not change and stayed at the default value of -1. As a result, the PVT motion that was generated from multiple motion modify's to append points was running out of frames.