Re: Motor Drivers for Applied Motion Products (Ethernet/IP)?

[Mark Rivers via Tech-talk <tech-talk at aps.anl.gov>]

One advantage of using the C++ approach is that you can use the EPICS motor record, which has fields to control all of the common motor parameters (velocity, acceleration, backlash, software limits, etc.) and functions (move absolute, move relative, jog, stop, home, etc.)

If you use the StreamDevice you need to create an EPICS database with records for all of these parameters and functions. You need to invent methods to implement more complex things like backlash, set position after homing, etc.

I think StreamDevice will be much more work in the end. There are C++ drivers for the motor record that are only a few hundred lines of relatively simple code, and you use these as a model.

Another advantage of the motor record is that all motors in your system appear logically identical, even if the control hardware is different. All of the OPI screens have already been written for you. These make replacing the controller with another type much easier in the future.

Mark

Sent from my iPhone

On Aug 3, 2024, at 11:24 PM, Feister, Scott via Tech-talk <tech-talk at aps.anl.gov> wrote:



Hi EPICS users,

 

Is anyone aware of work on drivers for Applied Motion Products? In particular, I’m looking at the “SSM23IP-3EG NEMA 23 Integrated StepSERVO Motor w/ EtherNet/IP”:

 

https://www.applied-motion.com/s/product/integrated-stepservossm23ip3eg/01t5i000000xzQjAAI?name=SSM23IP-3EG-NEMA-23-Integrated-StepSERVO-Motor-w-EtherNet-IP

 

If nobody has been working on a driver, I believe the correct approach for me would be to mimic the motor drivers at https://github.com/epics-motor. However, another temptation would be to use the StreamDevice toolkit as a faster pathway to the goal line. Can anyone speak to the pros and cons of a “StreamDevice” approach, versus writing a C++-code-level driver for a motor?

 

Thanks!

 

Scott