question for Motion control actual position

[angi]

hi guys

i need to know the Actual position of one Axis, do i need to use GVS to get the value or just directly use the value in the Axis.ActualPosition?

thanks

 

[Peter Nachtwey]

Use the Axis.ActualPosition. The ActualPosition should be returned every scan by default. GSV is for getting those items that aren't returned by default and the GSV is slower.

 

[MichaelG]

More info about how rockwell updates tags

http://www.plctalk.net/qanda/showpost.php?p=380967&postcount=23

 

[mallorz]

One thing that may or may not be worth mentioning is that the value stored in the .ActualPosition tag is delayed by one coarse update period.

 

[MichaelG]

One thing that may or may not be worth mentioning is that the value stored in the .ActualPosition tag is delayed by one coarse update period.

can you please explain this further

 

[Peter Nachtwey]

I am curious too. Remind me how this works.

can you please explain this further

Yes, mallorz, doesn't it make a difference if the position is coming over the back plane or over the Ethernet?

I modified and enhanced the M02AE code to become the HYD02 and M02AS.
It has been over 12 years since I worked on M02AS and HYD02 code and I haven't seen the code for over 10 years. I can find out about any delays over the back plane. My company still makes the HYD02 and M02AS for Rockwell so I can find out how the back plane modules work. I have no clue what happens over Ethernet.

 

[MichaelG]

Yes, mallorz, doesn't it make a difference if the position is coming over the back plane or over the Ethernet?

I modified and enhanced the M02AE code to become the HYD02 and M02AS.
It has been over 12 years since I worked on M02AS and HYD02 code and I haven't seen the code for over 10 years. I can find out about any delays over the back plane. My company still makes the HYD02 and M02AS for Rockwell so I can find out how the back plane modules work. I have no clue what happens over Ethernet.

I have the impression that the Ethernet behaves in this way:
The PLC sends regular packets (do not know if UDP or TCP or if they are marked with QoS) at the coarse update rate to drive
In each packet is a nano second timestamp followed by where the axis has to be over the next n (I got the impression that it was 3 to 4) coarse updates. For EACH Update the axis is told to be at X position with V velocity and A acceleration. This allows the ethernet to "miss" an update and the axis is still doing the correct thing.

I have not been told but I a$$ume that the reverse is happening from the axis

Note: the Whole system hinges on both devices (PLC and Axis) using EXACTLY the same time. Rockwell updated the ethernet timing system Version 18? ish and implemented a ethernet grandfather clock system utilising 1588 PTP (Precision time protocol) which syncs all devices to better than 1 ms (typically < 200ns)


Rockwell Blah Blah - ODVA Common Industrial Protocol Motion Brochure - CIP (ethernet) Motion
http://literature.rockwellautomation.com/idc/groups/literature/documents/br/00112-br001_-en-p.pdf

 

[Peter Nachtwey]

I have the impression that the Ethernet behaves in this way:
The PLC sends regular packets (do not know if UDP or TCP or if they are marked with QoS) at the coarse update rate to drive
In each packet is a nano second timestamp followed by where the axis has to be over the next n (I got the impression that it was 3 to 4) coarse updates.

That isn't good if you get a new command on-the-fly. The future can change.

For EACH Update the axis is told to be at X position with V velocity and A acceleration. This allows the ethernet to "miss" an update and the axis is still doing the correct thing.

This method doesn't allow for the missing of packets. It is the next best thing to the way I would do it. I would download the coefficients for the polynomial for the motion profile or a segment of the motion profile.

I have not been told but I a$$ume that the reverse is happening from the axis

???

Note: the Whole system hinges on both devices (PLC and Axis) using EXACTLY the same time. Rockwell updated the ethernet timing system Version 18? ish and implemented a ethernet grandfather clock system utilising 1588 PTP (Precision time protocol) which syncs all devices to better than 1 ms (typically < 200ns)

Yes.

We are members of the ODVA and have a person that monitors and attends the meetings. He would know the answer to this. I have not been too involved with this stuff because I have had my own thoughts of how all of this should be done and am waiting for the rest of the world to catch up.

BTW, I still think the best way to handle multiple axes is to gather calculate and output in parallel. This method minimizes phase delay between axes when one axes is gear to another.

 

[mallorz]

Read it in a manual last week. Here's an excerpt, it can be found on page 206.

"Actual Position is the current absolute position of an axis, in the configured Position Units of that axis, as read from the feedback transducer. Note, however, that this value is based on data reported to the controller as part of an ongoing synchronous data transfer process, which results in a delay of one coarse update period. Thus, the Actual Position value that is obtained is the actual position of the axis one coarse update period ago."

--SERCOS and Analog Motion Configuration and Statup Manual.
http://literature.rockwellautomation.com/idc/groups/literature/documents/um/motion-um001_-en-p.pdf

 

[MichaelG]

Read it in a manual last week. Here's an excerpt, it can be found on page 206.

"Actual Position is the current absolute position of an axis, in the configured Position Units of that axis, as read from the feedback transducer. Note, however, that this value is based on data reported to the controller as part of an ongoing synchronous data transfer process, which results in a delay of one coarse update period. Thus, the Actual Position value that is obtained is the actual position of the axis one coarse update period ago."

--SERCOS and Analog Motion Configuration and Statup Manual.http://literature.rockwellautomation.com/idc/groups/literature/documents/um/motion-um001_-en-p.pdf

yet again rtfm



An OVDA Paper presented by rockwell for CIP Motion ethernet transport

http://www.odva.org/Portals/0/Library/CIPConf_AGM2009/2009_CIP_Networks_Conference_Technical_Track_CIP_Motion_Implementation.pdf

 

[OkiePC]

yet again rtfm



An OVDA Paper presented by rockwell for CIP Motion ethernet transport

http://www.odva.org/Portals/0/Library/CIPConf_AGM2009/2009_CIP_Networks_Conference_Technical_Track_CIP_Motion_Implementation.pdf

Golden Nugget. Thanks.