Gap between probing and machine positions

[raoullector]

When I perform a probing operation in GrblHal using the G38.2 command, I noticed a discrepancy between the stopping position and the position recorded by the probe. This discrepancy increases with the probing speed.

For example, for a probing command G38.2 Z0.5 F500, the stopping position is:
<Idle|MPos:133.219, 0.000, 20.516 | Bf:100,1024 | F:0 | Pn:P>
Whereas the position recorded by the probe is:
[PRB:133.219, 0.000, 22.087:1].
Here, we observe a gap of more than 1.5 mm on the Z-axis.

However, when probing fails, both the machine and the probe report the same position:
<Alarm|MPos:133.219,0.000,0.502,0.000|Bf:100,1024|F:0>
And:
[PRB:133.219,0.000,0.502,0.000:0]

I understand that the machine requires some time to decelerate after the probe is triggered, and that this time depends on the initial speed.
But is the decel time the only reason for such a gap ?

PS: I use the jerk plugin.

[terjeio]

But is the decel time the only reason for such a gap ?

Yes. And the large gap could be due to the jerk option having a longer deceleration tail than normal. There is a known issue with jogging and this is, IMO, likely related.

FYI there is an option for emptying the segment buffer of cruising motion on probe asserted, this may help reducing the gap - but perhaps not much with jerk enabled, especially if a larger segment buffer is used. You can find it in the Advanced features tab in the Web Builder.

Perhaps jerk should be disabled when probing? @Dietz0r what do you think?

[Dietz0r]

Perhaps jerk should be disabled when probing? @Dietz0r what do you think?

Would this be possible without loads of coding that becomes obsolete the second the issue is fixed? as a workaround it's obviously a good idea but it seems like a lot of work? Maybe hardcoding probing accel/decel values?