Dear, collegues. I need your advice

TO_PositioningAxis does not have functionality to take into account curvature and backlash of mechanical transmission, so the position value from encoder will differ from the actual position of the carriage. Siemens Support provides function examples of backlash compensation:

https://support.industry.siemens.com/cs/ww/en/view/109766673

We can use compensated positions variables in our developed kinematic model. S7-1500T PLCs provides its own kinematicsl models in TO_Kinematics, to simplify a lot machine development. However, the TO_Kinematics technology object includes TO_PositioningAxis by default with its unchangeable variables.

Is there a way to add the corrected position variables to the TO_Kinematics technology object?

With regards, Maksim

There is always a way around a hardware-based limitation - if one exists.

If you can use M or DB tags instead of hardware inputs for the feedback address, by all means, do that.  However, if the TO block is restricting the value to hardware-based inputs, you may be able to alter the way the block reacts to the hardware input (aka, cheat).

By that I mean, if the TO blocks _only_ allow hardware addresses to parameterize a TO block, if it also allows you to specify the input range for scaling purposes, you can input the raw data into your backlash compensation model, and adjust the scaling of the input used by the TO block.

If this is not available, please post a screenshot of the TO's to see if there is something that you can change every call of the TO block that will produce the desired result.

There is also another way to cheat.

Access the commands from the TOs, and override them with logic that gets called right after the TOs complete their calculations.

I have seen people install analog outputs (or spare telegram data) not connected to anything!, and in the network after a fake output is updated by some block, that output is then modified and sent to the correct output.

Lots of ways to cheat the system.

Personally, I would prefer to virtualize these types of blocks, PIDs, TOs and such, but then the commissioning tools would probably not work right - in many cases.

MaximNevmer

Dear, collegues. I need your advice

TO_PositioningAxis does not have functionality to take into account curvature and backlash of mechanical transmission, so the position value from encoder will differ from the actual position of the carriage. Siemens Support provides function examples of backlash compensation:

https://support.industry.siemens.com/cs/document/109766673/simatic-s7-1500%3A-backlash-compensation?dti=0&lc=en-WW

We can use compensated positions variables in our developed kinematic model. S7-1500T PLCs provides its own kinematicsl models in TO_Kinematics, to simplify a lot machine development. However, the TO_Kinematics technology object includes TO_PositioningAxis by default with its unchangeable variables.

Is there a way to add the corrected position variables to the TO_Kinematics technology object?

With regards, Maksim

MaximNevmer

TO_PositioningAxis does not have functionality to take into account curvature and backlash of mechanical transmission, so the position value from encoder will differ from the actual position of the carriage.

Not really understanding the functionality of the thing I have two ideas to pass to you. If they can fit your needs - no idea though.

1) let's suppose you know exactly a displacement of a final element (a carriage?) as a function of a desired position, then you can introduce a virtual axis in place of the real one and then you can gear the real one to the virtual one by using a cam which compensates for the displacement,

2) you can use for positioning second encoder which reads exact position of a final element.

 Thank you. I think both solutions are pretty clear and applicable.

jacek d

 

MaximNevmer

TO_PositioningAxis does not have functionality to take into account curvature and backlash of mechanical transmission, so the position value from encoder will differ from the actual position of the carriage.

Not really understanding the functionality of the thing I have two ideas to pass to you. If they can fit your needs - no idea though.

1) let's suppose you know exactly a displacement of a final element (a carriage?) as a function of a desired position, then you can introduce a virtual axis in place of the real one and then you can gear the real one to the virtual one by using a cam which compensates for the displacement,

2) you can use for positioning second encoder which reads exact position of a final element.

I didn't understand your suggestions very well. Please clarify my reasoning. You suggest two ways to correct data into axis TO.

In the first method, you suggest "input the raw data into your backlash compensation model". I use telegram 105 IN, and can read encoder position in G1_XIST1 word, and write it as input of geometry compensation model. Using the backlash compensation function from Siemens Support in the MC-PreServo block, do I understand correctly that the function overwrites the values of G1_XIST1, and the position in TO already becomes corrected as a result of its calculations?

Am I right, that to realise first way I need to change Drive setting "Data connection" in TO Configuration to "Data block" (picture attached) and do all what needed with DB data?

In the second method (cheating) you suggest "access the commands from the TOs" - do you mean the words of telegram OUT, for example, G1_STW (Encoder), XERR (DSC position deviation)?

"and override them with logic that gets called right after the TOs complete their calculations" - do you suggest to add correction value to G1_STW value in MC-PreServo after calling the backlash compensation function?

0_Einstein

There is always a way around a hardware-based limitation - if one exists.

If you can use M or DB tags instead of hardware inputs for the feedback address, by all means, do that.  However, if the TO block is restricting the value to hardware-based inputs, you may be able to alter the way the block reacts to the hardware input (aka, cheat).

By that I mean, if the TO blocks _only_ allow hardware addresses to parameterize a TO block, if it also allows you to specify the input range for scaling purposes, you can input the raw data into your backlash compensation model, and adjust the scaling of the input used by the TO block.

If this is not available, please post a screenshot of the TO's to see if there is something that you can change every call of the TO block that will produce the desired result.

There is also another way to cheat.

Access the commands from the TOs, and override them with logic that gets called right after the TOs complete their calculations.

I have seen people install analog outputs (or spare telegram data) not connected to anything!, and in the network after a fake output is updated by some block, that output is then modified and sent to the correct output.

Lots of ways to cheat the system.

Personally, I would prefer to virtualize these types of blocks, PIDs, TOs and such, but then the commissioning tools would probably not work right - in many cases.

Hello,

further possibility.

It is possible to use datablock for the interface of the axis and copy the telegram data in MC_Preservo/Postservo. Then it is possible to manipulate the raw position raw value.

Using the MC-PreServo and MC-PostServo organization blocks

https://support.industry.siemens.com/cs/ww/en/view/109741575

Thank you, that is really helpful. 
Can you advice manual with description of all telegrams?

Technology Team

Hello,

further possibility.

It is possible to use datablock for the interface of the axis and copy the telegram data in MC_Preservo/Postservo. Then it is possible to manipulate the raw position raw value.

Using the MC-PreServo and MC-PostServo organization blocks

https://support.industry.siemens.com/cs/ww/en/view/109741575

Hello,

3.9 PROFIdrive

https://support.industry.siemens.com/cs/ww/en/view/109781807