Toolchanger: print- and toolchange acceleration settings

[endstop]

My Orca-slicer generated GCODE is full of "SET_VELOCITY_LIMT ACCEL=... ACCEL_TO_DECEL=...".
I just noticed on an aborted print that those may carry over to the toolchange sequence, which seems unintended. I had slowed down the first layer to a crawl, but the next tool change happened in slow-motion as well.

Is there maybe some documentation, insights, general strategy etc. how to address toolchange-vs-print limits management?
I believe I can hack away the problem e.g. restore settings in my toolchange macros. But I would be surprised if I'm the first one to run into this.

 

[vinnycordeiro]

This is the expected behavior on any firmware: you make a change on the speed, the change is kept until the printer is told otherwise. You need to modify your settings to increase the speed after the lower one isn't needed anymore.

 

[endstop]

Thanks - I understand that for "G1 Fxyz". KlipperToolchangerEasy overwrites this slot in the printer state , probably even saves and restores GCODE_STATE being a good citizen.
But "SET_VELOCITY_LIMIT" appears to be a different "slot" in the printer state that orca slicer does not restore and KlipperToolchangerEasy seems to be fully unaware of (with the exception of one usermods file)?

 

[endstop]

...maybe it is just machine humor when it picks up the toolhead with exaggerated care for layer two, then goes into a twitching spree

 

[endstop]

Update: I just ran into a situation where the Z steppers (Voron 2.4) skipped steps during toolchange. Passed my own kinematics test hitting SHIFT-CTRL-E halfway across the room.
It seems a fundamental problem now (one that is easily solved, though).

There is an inherent trade-off between sustainable max. acceleration and max. speed (stepper drive voltage opposed by back-EMV and inductance, which both scale proportionally with v). Apparently, Orca slicer sets both. They are consistent, or violating the speed-/accel envelope would cause obvious layer shift already while printing layers.
But combine Orca's acceleration setting with the speed I've tuned for the toolchanger (with lower acceleration - as it travels 2x 300 mm for every change, speed not accel is the bottleneck) and the Z motors run out of torque during tool change.

The writing is on the wall - I have to set my own "tool changer" acceleration before tool change that is sustainable with the tool change speed I allow for KTE. This may look obvious and it's trivially done. Still I'm wondering, am I the first one to run into (or at least "to document" this feature gap)?