Firmware pressure control

Hi. Im fine tunning my printer, prusa i3, direct extruder.
I have fat corners. I have been readon this is a normal problem caused by acceleraton, speed and extruder pressure not catching up.
Also I  read pressure control is used to imprve this. I have been searching but all info is not really clear to me, so I need some help please:

1. I have found this in configuration:
#define EXT0_ADVANCE_K 0
#define EXT0_ADVANCE_L 0
#define USE_ADVANCE 0

I have read that you do not recommend quadratic. For a direct extruder (mine is Greg's extruder type if it matters and well calibrated) wich values I have to tune? The K or the L value?

2.- I have read that I can set the value through gcode command wich is great because I guess avery filament would have its own values. In order to be able to set the values through gcode (M233 command?) do I have to set #define USE_ADVANCE to 1 or should I leave it to CERO?

3.- I couldnt find the geenrator you talked about here:
Useing that generator I could get values for K or L (wich value is the one I have to tune?)

4.- So I can use this method to calibrate L?: that would give me L, but what about K?

5.- I cant find the M233 command in the list of gcode commands, is that right?

Thanks in advance for the help


  • 1. L value, 40 for direct extruder is a good start.
    2. YOu need USE_ADVANCE to able to enable it at all. Then as soon as L is non 0 it is active.
    3. That ha sbeen removed from host. Better use 4
    4. yes just replace commands for setting L
    5. See repetier.ino in firmware it explains all implemented gcodes.
  • edited April 2018
    We only have to tune the L parameter we dont touch the K?
    Also, I use slic3r: I have retraction turn off, also no retraction in formware. Im printing a hollow cube to test how the corners comes. I have tried L 40,50 and 60 I dont see improvements (what I see are corners bulkie, a common issue). What Im observeing too is that I use a skir in order to prime the noozle. When print starts, the extruders gets crazy: It overextrudes a LOT at each coerner and underestrude in the lines of the sides of the cube. Then I see a lot of new activity while priting the solid bottom of the cube. The after that the walls are printed, I can see something is different in the extruder way of moveing but cant see a real difference in the corners. Wich range of values the L factor takes, and what about K? should I set the K value too?
  • K is quadratic term and best left to 0 or extruder will run like crazy on faster moves.
    Do not overestimate the result achievable. Best is to use the marlin test path where you can really see the difference just as they describe. Then you know best possible result.

    With bigger L/K the amount extruder has to do extra increases and with high acceleration it gets even worse and can result in even reverse extrusion direction. That is where it gets unusable normally.
  • And what values you recomend for acceleration. I have 700 and Im going to check but I think I have tried with 400 too,
  • The printer where I tested has even 1100 which is no problem with L=40 and max print speed 70-80. Accelerations of 2000-7000 would be really high for advance.
  • doing the calibration pattern, Im useing a value of L+40 . In the pattern it looks god. The inserting the command M233 Y40 into my gcode of a print I see 2 things:
    1.- corner of the brim is FAT, way way much than before and then the lines are heavily underextruded. Things go much better durein the rest of print.
    2- dureing the print I hear like if the extruder steper motor is looseing steps.

    What you think about this?

  • By the way, Im useing version 0.92
  • 0.92 might be the problem. I know I fixed some issues on this in 1.0.x.
Sign In or Register to comment.