`spindle-at-speed` is not waited for when disabling the spindle

[wucke13]

I do have something like this as G-Code:

M5
G91
G1 Z+0.001
G38.3 Z-0.5

regardless of what the spindle0.at-speed pin gets as signal, the moves after M5 begin immediately (also regardless of whether the G1 is there or not). I would like to avoid having a spinning tool hit my toolprobe; so I think there is legitimate reason to also consider at-speed when the spindle is off and or the commanded spindle speed is 0.

Actually I would expect G38.n to form a at-speed barrier by itself, without needing to resort to the hacky G1 above to wait for the spindle to be at speed. Attached is a halscope log from the following actions:

• Enable spindle
• Start the G-Code listed above

As can be seen, the Z movement starts immediately after the M5, despite at-speed lagging almost 10 seconds behind.

halscope.csv

[andypugh]

I don't know if it will help, but can you log motion.program-line?

Spindle-at-speed does go to true, that may be enough to trigger motion.
As far as I know spindle-at-speed going false does not halt motion (I don't think that you are expecting it to, though?)

Maybe there should be a built-in behaviour to actually block probing if the spindle-speed-in is non-zero.

[grandixximo]

There are laser probes that support probing with spinning spindles, preventing probing at spindle-speed = 0 should be opt in at best IMO.

[wucke13]

As far as I know spindle-at-speed going false does not halt motion (I don't think that you are expecting it to, though?)

Yes, you get me right. What I expect is, that spindle-at-speed must be true before the next G1/G38.n/... is started. And of course, that M5 (effectively commanding the spindle speed to be 0), is treated like any other spindle speed.

Maybe there should be a built-in behaviour to actually block probing if the spindle-speed-in is non-zero.

There are laser probes that support probing with spinning spindles, preventing probing at spindle-speed = 0 should be opt in at best IMO.

Both of which is good reason to not change G38.n to react to the set spindle speed. I do think however, that measuring should only start when the spindle is at speed.

To construct a rather weird argument: When probing a large one-flute endmill via laser, it might be necessary to have the spindle turning so that the lowet point of the endmill is properly detected. In this case one would command a spindle speed, and only then G38.n. The G38.n should wait until the spindle is up to speed before probing.

Equally (as laid out above already), if the spindle speed is commanded to be 0 (i.e. via M5), than the G38.n shall wait until the spindle reached this speed.

[grandixximo]

Yes, that is the correct expectation, and how it should work, the current behavior is wrong.

[grandixximo]

@wucke13 could you test the PR please?

[wucke13]

Will do tonight

[wucke13]

Feedback is here: #4160 (comment)