I suspect this is related to a problem that was discussed on the developers' list in July: https://www.mail-archive.com/search?l=mid&q=20160721150035.GA46232%40unpythonic.net
Basically, we noticed that the charge-pump pin would be incorrectly activated "for a huge fraction of a second" during startup, when it should only be activated when the machine is taken out of estop. The cause turned out to be the way the "net" (and "link") commands can set a default value for a signal, coupled with the fact that it takes a long time at startup before pins like iocontrol.0.user-enable-out get driven with their correct value. We have implemented a fix in master branch only (so the fix will first be in version 2.8.0, whenever that is ready), as discussed in that thread. https://github.com/LinuxCNC/linuxcnc/commit/106fcd8a02b569b7ee8c017af7905856a704535b If the initial value (before halcmd 'start') of the *writer* pin on the signal is correct, then one workaround is to make sure that you net/link the writer pin *first*. For example, the original problem can be fixed by changing net estop-out charge-pump.enable iocontrol.0.user-enable-out to net estop-out iocontrol.0.user-enable-out charge-pump.enable or net estop-out iocontrol.0.user-enable-out net estop-out charge-pump.enable Otherwise, the workaround would be to link any reader pin, then explicitly set a value onto the signal, and then link the rest of the signals to the pin: net estop-out charge-pump.enable sets estop-out false net estop-out iocontrol.0.user-enable-out In a future version with the fix, this workaround should be unnecessary but harmless. Jeff ------------------------------------------------------------------------------ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
