Dear Smoothie community.
Jim Morris just merged a massive code refactor into Smoothie ( https://github.com/Smoothieware/Smoothieware/pull/961 ).
Thanks to him for this truly massive chunk of work, and to all those who tested it before merging.
This changes Smoothie to calculate acceleration for every step instead of 1000 times a second, which results in even smoother/better quality movement.
It also comes with many many smaller improvements to planning and step generation. It solves many small problems users experienced in some circumstances. Amongst those, it is expected to fix the infamous “S3D bug”, so please test that.
It is also expected to make Smoothie wore sturdy overall, and is a first step towards implementing some more refactors.
A few things have changed for users, so please read https://github.com/Smoothieware/Smoothieware/blob/edge/upgrade-notes.md before upgrading.
To upgrade, the procedure is as usual, and you can find it here : http://smoothieware.org/flashing-smoothie-firmware
Please tell us if you encounter any problems, and what you think about your Smoothie’s new step generation.
Cheers 
PS : The code does not support laser yet, do not upgrade to this new version if you are using a laser. We will implement laser as soon as possible.
Imported from wikidot
Bonjour à tous
Désolé en français, j’ai chargé le nouveau soft et modifié que la direction de l’extruder dans le fichier config. Je constate sur mon imprimante cartésienne à base de EShapeoko que le démarrage et la prise de position zéro semble beaucoup plus fluide sans à coups, ensuite il est difficile de voir la différence mais je suppose qu’il doit y en avoir une comme pour le démarrage, quand au résultat sur un objet simple je ne vois pas de changement du moins sur l’aspect, on continue avec cette version et si je ferais part de mes constatations.
Bonne journée à tous
Super, merci du retour.
Great, thanks for the feedback
Steps missing or miscalculated on a delta printer
Tested printing with the newest edge branch firmware, the part I print kept tilting/drifting towards Z axis. Switched 3 physical axis(X to Y, Y to Z, Z to X) but the problem still comes as drifting towards logical Z axis.
What’s more, the drift seems kept a certain distance each Z layer. For example, when printing things of the same height, printing at 0.1 mm layer height seems drifts 3 times the distance of the situation when printing at 0.3 mm layer height.
G code test results:
Fine with code like
G0/G1 Z300
G0/G1 Z5
Problems comes up with:
repeating
G90
G1 X20
G1 Y20
G1 X-20
G1 Y-20
G91
G1 Z0.1
drift/tilt towards Z axis at each Z level.
BTW, when speed up to a certain level (maybe reached the max pulse frequency), it works fine. Sometimes I tested it at over 500% speed, and it works fine, but the problem comes up at the speed below 500%, even at 495%.