• Tags
  • Documentation
  • Order
  • Register
  • Login
Duet3D Logo Duet3D
  • Tags
  • Documentation
  • Order
  • Register
  • Login

What are the units of jerk?

Scheduled Pinned Locked Moved
Tuning and tweaking
6
15
6.1k
Loading More Posts
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • undefined
    peridot
    last edited by 24 Sept 2016, 00:43

    Just to check, the units of jerk in M566 are millimeters per minute, not per second, right? As are the maximum speeds? But the accelerations are millimeters per second per second?

    I'm trying to diagnose some clunking noises and overshoots, and I want to make sure I know what the values I'm feeding to jerk mean.

    1 Reply Last reply Reply Quote 0
    • undefined
      curieos
      last edited by 24 Sept 2016, 00:58

      I would assume jerk is in mm/s^3 or mm/m^3 (as its the derivative of acceleration)

      She/Her
      I work at a local 3D printing shop.
      Printers: Micron+ w/Duet 3 Mini, in-progress adaptation of the Jubilee REL onto an E3D MS, Prusa i3 MK3S.

      1 Reply Last reply Reply Quote 0
      • undefined
        peridot
        last edited by 24 Sept 2016, 01:47

        "Jerk" as used in 3D printers is sort of a misnomer. It's conventionally used to mean "the largest instantaneous change in velocity that it is safe to command", so it is in velocity units. Just, we seem to be stuck on a mix of mm/s and mm/min. And since I don't have any clear a priori idea how big a jerk is realistic, it's not obvious from the values people quote which.

        1 Reply Last reply Reply Quote 0
        • undefined
          dc42 administrators
          last edited by 24 Sept 2016, 07:01

          In the M566 command supported by RRF the units are mm/min as for all other speeds.

          Duet WiFi hardware designer and firmware engineer
          Please do not ask me for Duet support via PM or email, use the forum
          http://www.escher3d.com, https://miscsolutions.wordpress.com

          1 Reply Last reply Reply Quote 0
          • undefined
            Hugues
            last edited by 24 Sept 2016, 08:50

            Is it possible to use macro that can store different profile for acceleration, jerk and motor current, safe and agressive and can be launch while printing ?
            Just an idea and i'm not near my printer…

            1 Reply Last reply Reply Quote 0
            • undefined
              peridot
              last edited by 24 Sept 2016, 09:05

              That should be no problem: create a macro, a g-code file that sets the settings you want (aggressive.g, safe.g), and you can run it from a button on the printer page or use "M98 Paggressive.g" in the start g-code for a particular print. See http://reprap.org/wiki/G-code#M98:_Call_Macro.2FSubprogram.

              1 Reply Last reply Reply Quote 0
              • undefined
                Hugues
                last edited by 24 Sept 2016, 11:29

                Nice, i've to setup this to make some tries. I'm experiencing skipping step where there is no difficulties (circle path).

                1 Reply Last reply Reply Quote 0
                • undefined
                  bot
                  last edited by 24 Sept 2016, 16:54

                  Just as a note on jerk values, I find that a much lower value than is typical gives great results if you're doing lots of tiny intricate details. I use 480 mm/min. With higher values, such as 900 or 1200, my machine didn't sound happy when I commanded lots of tiny direction changes.

                  *not actually a robot

                  1 Reply Last reply Reply Quote 0
                  • undefined
                    peridot
                    last edited by 24 Sept 2016, 18:03

                    What is the downside of using a really low value for jerk? With an acceleration of 3000 mm/s/s, changing a jerk of 300 mm/s to zero just adds a tenth of a second (per occurrence) to the print time.

                    1 Reply Last reply Reply Quote 0
                    • undefined
                      bot
                      last edited by 24 Sept 2016, 19:01

                      The print time does drastically go up, depending on how many times the toolpath changes directions quickly. For me, it's an increase of about 30% of print time, minimum (going from 900 to 480 mm/min jerk). I'm doing quite complex prints, though.

                      Also, with low jerk values, you encounter problems with the varying speed of extrusion causing blobs on corners, something that M572 works to address. Be careful, though, because there appear to be some infrequent bugs in the M572 code when small values (below 0.05) are used.

                      *not actually a robot

                      1 Reply Last reply Reply Quote 0
                      • undefined
                        Hugues
                        last edited by 24 Sept 2016, 20:19

                        Also, i noticed that increasing acc/jerk for extruder is really helpfull to get better print. I use a e3d titan and i setup it to the same value as xy motion.
                        About M572 bug i use s 0.05 and it works great for my printer.

                        1 Reply Last reply Reply Quote 0
                        • undefined
                          roboduet
                          last edited by 24 Sept 2016, 22:24

                          What settings (accel. and jerk) do you use for geared extruders? I'm using ~5.18:1 planetary gearbox with 12 mm OD spur gear, 450 steps per mm (1/16 microstepping).

                          @cubexupgrade:

                          About M572 bug i use s 0.05 and it works great for my printer.

                          I'm using ~600mm bowden tube and tried this M572 option with different parameters. I hear other sounds from my extruder when I using this option, but I didn't noticed any visible differences in print quality. Maybe I'm doing something wrong?

                          1 Reply Last reply Reply Quote 0
                          • undefined
                            Hugues
                            last edited by 25 Sept 2016, 06:57

                            I set 5000 in acceleration and 2000 in jerk for pla but for ABS it's lower (filament is been grinding), i didn't try higher value for PLA.

                            M572 helps to avoid stringing for me and sharp corner, i' ve a small bowden tube like 30 cm. For 60cm you should use higher value and try to find your higher value for extruder.

                            1 Reply Last reply Reply Quote 0
                            • undefined
                              bot
                              last edited by 25 Sept 2016, 16:43

                              @cubexupgrade:

                              Also, i noticed that increasing acc/jerk for extruder is really helpfull to get better print. I use a e3d titan and i setup it to the same value as xy motion.
                              About M572 bug i use s 0.05 and it works great for my printer.

                              I don't think it's typical to get better print quality with higher accel/jerk values. It is almost always the other way around: better print quality is achieved with lower values.

                              There have been, in the past, some issues with running lower jerk values, but I think David worked out some bugs in respect to that. IMO, the LOWEST accel/jerk/feedrate settings should be used that still provide adequate print speed – this will allow the least ringing artifacts, least likelihood for motor stalling, and least unwanted vibrations.

                              *not actually a robot

                              1 Reply Last reply Reply Quote 0
                              • undefined
                                Hugues
                                last edited by 25 Sept 2016, 19:46

                                I agree with this before, but i watched closely the dice project from Rene Jurack where he tested lot of setup to finish with high value. I made my tests in this way.
                                For the extruder high value i cannot explain clearly why they gave me better result…

                                1 Reply Last reply Reply Quote 0
                                7 out of 15
                                • First post
                                  7/15
                                  Last post
                                Unless otherwise noted, all forum content is licensed under CC-BY-SA