dls-controls / ethercat

EPICS support to read/write to ethercat based hardware

ethercat

EPICS support to read/write to ethercat based hardware

Prerequisites: IgH EtherCAT Master for Linux

This EPICS module builds with a patched version of etherlab, described in the file etc/makeDocumentation/building.src

The documentation was made when doxygen at DLS would build in the folder etc/makeDocumentation.

The doxygen documentation is no longer building at Diamond, but the "sources" are in the folder etc/makeDocumentation.

Release notes in etc/makeDocumentation/release_notes.src

Please email with issues as the maintainer has been know to ignore github notifications for months.

Maintainer notes

Before making a release:

1. Check/update the version in ethercatApp/scannerSrc/version.h
2. Write release notes in etc/makeDocumentation/release_notes.src

ethercat's Issues

string needs to be unpacked

https://github.com/dls-controls/ethercat/blob/master/ethercatApp/src/ecAsyn.cpp

The method receive_config_on_connect has debugging messages that attempt to print the configuration received from the scanner.
printf is called on the config_buffer "raw". However the scanner sends the configuration packed and the message includes the length of the following string.
So the output is garbled. Example output from ioc running version 4-3

[Fri Nov 13 18:47:58 2015]getting config
[Fri Nov 13 18:47:58 2015]config-file size:111970
[Fri Nov 13 18:47:58 2015]^D
[Fri Nov 13 18:47:58 2015]************************

how to set for 1262 2262 3742 4712 .. over sampling modules

such terminals need DC for sampling sync. How to set scanner for such slaves to use one reference clock of a slave terminal or the master clock? We could not get the sample values if did not set the sync unit clock.

Missing git tags?

The last release shown on github is version 6-1, but I just noticed that your commits and release notes go all the way up to 6-5. I suspect you may have forgotten to git push --tags

Getting invalid data into EPICS

When we run the scanner with 1khz (1ms) we sometimes get invalid data into the EPICS layer. This is largely solved by lowering the frequency to 667hz (1.5ms) but in theory the issue remains.

I think that invalid data never should be propagated to the EPICS layer.

What do you think? Do you need any more specific logs/data?

cc: @alex-soderqvist

build instructions for patched etherlab

The documentation needs updating because
etherlab migrated to git at gitlab
https://gitlab.com/etherlab.org/ethercat and the instructions to build the module in etc/makeDocumentation/building.src still refer to mercurial