Pressure advance calibration, the speed maxes out at 25mm/s



  • So was doing two pressure advance tests today, one from 0.35 to 1.10 and another from 1.10 to 3.10. I set the print speed to 80mm/s, and I noticed that the interface said it was actually going about 22 to 25mm/s at maximum.

    I am doing 2 perimeters, and it ended up doing the external perimeter at 40mm/s (50%) - although actual speed was again 22 to 25mm/s. It is using the script provided on the documentation page that describes a calibration.

    I don't know if it was because of the crazy slow speed or not, but the ridge on the z seam doesn't really change. But I started noticing patterns in the second test's walls (they wouldn't be visible to a camera because of dark filament and low light). They feel like sections of the walls have slight thickness variations almost causing a wave feel.

    I noticed the side effect of enabling pressure advance indicated it might slow down because of jerk values. When I recently upgraded to RRF3 I might have made some errors perhaps? Opinions on my configuration would be valued: https://gist.github.com/madebyjeffrey/883d9b626b647376cd3134a3569f2777

    My machine is a mini kossel upgraded to have a duet wifi, linear rails, smart effector, v6, and a 600mm bowden tube on a bondtech extruder. Current filament is PLA.

    EDIT: I wanted to add a note. I noticed the speed decrease explicitly towards the end when the PA was higher, I don't remember what it did prior.


  • Moderator

    3.1 is a very high value for PA.

    You may need to increase your E jerk and E acceleration. Try 3000 or more for each.

    Also, check that you don't have a minimum layer time set in the slicer that could be artificially slowing down the entire print.



  • I just checked the values that were actually active and got these numbers:

    Maximum jerk rates (mm/min): X: 1200.0, Y: 1200.0, Z: 1200.0, E: 40.0, jerk policy: 1

    Accelerations (mm/sec^2): X: 3000.0, Y: 3000.0, Z: 3000.0, E: 120.0

    It is like it is being set some place but I haven't seen them be set anywhere else in the config file. I will have to go through the other files.

    It might explain why certain things aren't working. When I tried setting them particularly higher (which turns out to be massively higher) harsh noises were made.



  • It was in homedelta.g

    Is there a way to have this thing save some settings like saving them onto a stack, changing them, and popping them off after?


  • Moderator

    I don't know why you'd have jerk and acceleration values for the extruder in homedelta.

    If you need to change a value temporarily, just send the modified value, do what you need to do, and then set it back again. No stack needed.

    If you want to set a default value often and in many places you can use a macro for that. Have the command you want in the macro and then you can call it from other macros with M98 Pmacroname.g



  • It was generated by the configuration. I removed the E values from there.

    Now after I set:

    M566 X1200.00 Y1200.00 Z1200.00 E200.00 P1
    M201 X3000.00 Y3000.00 Z3000.00 E500.00

    For the range of PA values 0.5 to 2.5, something interesting happened.

    Above S0.5, there is a lot more vibration. At no place does the seam reduce much, it is always there. Starting at S1.25 the corners start to deteriorate, except for the one immediately after the seam. It definitely just looks as if it stops extruding in ever increasing distances from the corners affected.


  • Moderator

    200 and 500 is still an order of magnitude too low.



  • When I was using thousands the extruder was making a lot of noise. I can study it a little more tomorrow to find out what values it happens at.


  • Moderator

    There will be some more noise from the extruder as it pulls back rapidly.

    It would help to get some more details about your setup. Config.g, retraction settings, extruder type.



  • I posted my config: https://gist.github.com/madebyjeffrey/883d9b626b647376cd3134a3569f2777

    The acceleration and jerk values for the extruder are not totally up to date. Should be the 200 and 500 mentioned above.

    Retraction presently is 2mm @ 40mm/s - set in slicer.

    Extruder is a bondtech connected to a 600mm bowden tube to the smart extruder with v6 and a .35 nozzle attached.



  • I have finally gotten some results with acceleration of 3000, and jerk of 400.

    Any higher of a jerk and it makes too much noise, 400 still has some noise but it is almost reasonable.

    In my test I had the corners progressively get worse, and finally the Z seam.

    At no point did the Z seam reduce amount.

    I figure a good value is roughly in the 0.5 to 0.8 range, and I am testing that now. The documentation seems to suggest 0.5 is a reasonable place to be, which looks right. Although it is questionable if there will be a way to really see much difference here.

    I am also looking for other strategies to reduce the z seam as well.


  • Moderator

    I'd experiment with faster retraction speed as well.



  • I cannot say that it is perfectly calibrated, but I have s3d and prusaslicer producing reasonably good models on some basic calibrations. So this might be sufficient.


  • Moderator

    Some days that's the most I can hope for.


Log in to reply