M62/M63 vs M64/M65

[phil-barrett]

The LinuxCNC documentation on M62/M63 says

The actual change of the specified outputs will happen at the beginning of the next motion command. If there is no subsequent motion command, the queued output changes won’t happen. It’s best to always program a motion G-code (G0, G1, etc) right after the M62/63.

It appears that actual motion has to happen for the M62/M63 command to occur. If you use a G1 to the current location (i.e., no actual move), it fails to execute the M62 or M63. I have an application where I drive X/Y to a location, wait for a settling period of time and then trigger a digital output to start a somewhat slow event (sensor sampling). Motion should not happen during the sampling period. Unfortunately, motion is required to cause the M62 command to be executed. This may well be as designed.

The solution is to use M64/M65 and a synchronous sender (like ioSender with agressive (sic) buffering turned off). M62/M63 won't work for this application with agressive buffering. Many (most?) senders do not have an option to turn it off.

[terjeio]

G4P0.1 to complete buffered motion before M64/M65 will also work?