Synchronized Motion Notes	BKT 10-Nov-03

The term "synchronized motion" refers to the capability for one table controller to provide the reference inputs to another table controller, in addition to providing its own.  Therefore a single reference source provides reference inputs to both tables simultaneously, assuring synchronized motion.  Both table controllers have this capability, i.e., Table 1 can drive Table 2, and vice versa.  In addition, xPC Target can drive both tables and the structural test controller as well.

The term "master" refers to the driving table and the term "slave" refers to the driven table.  Whether or not a controller is a master or slave is determined by the selection of program source on the main panel:  If the program source is xPC Target or another table, the the controller is a slave, otherwise it is a master.  For independent motion, both controllers are masters.

When using this capability, there are several subtleties and caveats that you must be aware of to avoid surprises and test specimen damage.  These are discussed below: 


1) When viewing the "reference" testpoints of two channels in the Digital Oscilloscope or the Data Recorder, it may appear that there is a time delay between channels, even though the two channels are supposed to be perfectly synchronized.  This delay is not "real" from a control standpoint.  The reference signals of all channels pass through a delay line before being displayed so they can be aligned with their respective feedbacks.  THIS IS FOR DISPLAY AND ERROR LIMIT DETECTION PURPOSES ONLY.  The servocontrollers use the undelayed reference signals for command inputs, not the delayed versions.

To complicate matters, the values of delay that the master applies to the reference signals depends on whether a particular reference is intended for the master or the slave, whether or not Adaptive Control is active, and on which adaptive controller is active.  Below is a chart that shows what delay is applied by whom under what circumstances:


Adaptive	Delay Applied by Master		Delay Applied by Master		Delay Applied by Slave
Controller	to Master references		to Slave references		to Slave references

None		TVC "Input-output delay"	0				TVC "Input-output delay"
APC		0				0				0
AHC		TVC "Input-output delay"	0				TVC "Input-output delay"
AIC		AIC "Anticipation"		0				AIC "Anticipation"
OLI		0				0				0

The "Input-output delay" is an adjustment on the TVC panel and varies according to channel.  The "Anticipation" delay is an adjustment on the AIC panel and applies to all channels.  Note that the master applies no delay to reference signals intended for the slave; however, once these references reach the slave, the slave itself applies delay according to the above chart.

If the above information seems too complicated to consider, then perhaps is it better just to simply trust that the references are indeed perfectly synchronized, however skewed they may appear.  If you would prefer not to trust but to verify that this is so, turn off all adaptive controllers and set the "Input-output delay" of all channels in the TVC panel to zero, and then observe the reference signals in the Digital Oscilloscope.

Finally, it should be noted that the master writes all reference signals, whether master or slave, onto Scramnet in undelayed form for purposes of external data acquisition.


2) Even though the reference signals are perfectly synchronized, it may be that Feedback signals are not, due to input-output delay differences between controllers.  This kind of time skew has several causes:

a) When AIC in active in one controller but not the other, a potentially large time skew can occur.  AIC delays the reference by an amount of delay set by the "Anticipation" delay adjustment on the AIC panel, which can be many seconds long.  So if AIC is on in one controller but not the other, their feedbacks will be out of sync by an amount equal to the anticipation delay.  This could result in damaging forces if a stiff test specimen is spanning the two tables.  Therefore, when using AIC, make sure to turn it on in both controllers, and use the same "Anticipation" delay in both controllers.

b) Differences in OTM/OCL "Lead" adjustment between master and slave controllers can cause time skew, because the OTM/OCL compensators provide lead by delaying the reference signals.  The amount of lead used is typically small and so the resulting time skew has little if any detrimental effect.

c) The sample clocks of the master and slave controllers are not synchronized.  Depending on the phase difference between the two, the reference signals sent to the slave by the master may be one sample (about 1 millisecond) early or late.  This tiny amount of time skew has little if any detrimental effect.


3) Adaptive controllers APC, AHC, and AIC work in either master or slave mode.  OLI works only in master mode.


4) In Preview mode, reference signals can be previewed on the Digital Oscilloscope for verification without system motion.  On a master controller, motion is precluded by disconnecting master reference signals from the servocontroller command inputs.  However, this is not done for slave reference signals, which are never disconnected from Scramnet.  This will cause unexpected motion on the slave table if it is in "Run" mode.  Therefore, remember that "Preview" mode on the master only applies to the master, not to the slave.  If you want to preview the slave references, use "Preview" mode on the slave.

A good strategy to prevent unexpected slave table motion is to develop a habit of examining the "External ref" testpoints on the Digital Oscilloscope prior to hitting the Run button.  These testpoints show the current state of the slave reference input signals.


5) On both master and slave controllers, the start of data recording can be synchronized with the start of motion by checking the "Synchronize with Run" checkbox and hitting the "Arm" button in the Data Recorder Panel of both controllers.  Data recording on both controllers will then start the moment the master's Run button is pressed (not the slave's Run button).  After the test, the master and slave data recorder files can be merged using a third party data analysis tool.  Keep in mind however that the data in the two files may be time skewed by one sample (about 1 millisecond) due to the effect described in Item 2c above.


6) Always remember to hit Run on the slave first, then hit Run on the Master.  To stop, do in the opposite order, i.e., hit Stop on the master first, then hit Stop on the slave.
