search

InsanityAutomation / Marlin

Optimized firmware for RepRap 3D printers based on the Arduino platform.
http://www.marlinfw.org/
GNU General Public License v3.0
448 stars 220 forks source link

Quality issue with DW6 2.0.5 #134

Closed daredevil-ferreira closed 4 years ago

daredevil-ferreira commented 4 years ago

Description

I updated both motherboard & screen firmware (hex file and DWIN_SET provided on github)

And try to print these parts : https://drive.google.com/open?id=14RO5IvE9s4Z2HT-RN2ahGumINT-ekl8X 20200504_115548

with 2 different filaments.

I have a CR10s Pro and launch these prints via Octoprint.

The round shapes are ugly with a lot of blobs where as the length are more or less ok.

I noticed that the small movement are not continuous as if the motherboard was not able to execute smoothly the gcode.

Is it the Octoprint interface of the fact than now the screen displays every movement (X, Y, Z) while printing ?

I am going to test it directly from an SDcard.

chunter1 commented 4 years ago

I noticed that the small movement are not continuous as if the motherboard was not able to execute smoothly the gcode.

I observed the same after updating a few days agoe on my CR-10S Pro. However, i can't tell if the new firmware is less performant in executing micro movements. Maybe temporarely disabling the LCD interface feature to check any performance issues could help? Can you provide the link or gcode of your model for further investigations?

daredevil-ferreira commented 4 years ago

My GCode https://drive.google.com/open?id=10i5Plwj045Gf7a5AVMZxOFtqspzT7BaU

CCR10SPRO_lid (1).gcode.txt

How can I disable the LCD ?

chunter1 commented 4 years ago

How can I disable the LCD ?

Don't know exactly how but maybe by disabling the "EXTENSIBLE_UI":

// // Third-party or vendor-customized controller interfaces. // Sources should be installed in 'src/lcd/extensible_ui'. //

if ANY(MachineCR10SPro, MachineCRX, MachineEnder5Plus, MachineCR10Max) && (DISABLED(GraphicLCD) || ANY(Force10SProDisplay, ForceCRXDisplay))

define EXTENSIBLE_UI

endif

FrancescoTalotta commented 4 years ago

Hello. I noticed the same. Flashed the firmware 5 days ago and I have a lot of bubbles in the curved surfaces. I have a CR10s Pro with Octoprint. Do you know a commit that I can checkout in order to get rid of this problem?

image

riri19 commented 4 years ago

Same for me :( 20200504_191321

daredevil-ferreira commented 4 years ago

I had the same conclusion between SD & Octoprint SD: 20200505_121630

I did suspect an issue with the baud rate and I swith my raspberry 3B to a raspberry 4B and things get a bit better but still not as good as the SD print. Raspberry 4 vs Raspberry 3: 20200505_124726

Raspberry 4 vs SD: 20200505_124651

I am printing form the SD with same filament to verify. If there is still a reference, I will try to change the firmware for 115000 bauds.

riri19 commented 4 years ago

I hope you find a solution :)

Le mar. 5 mai 2020 à 1:02 PM, daredevil-ferreira notifications@github.com a écrit :

I had the same conclusion between SD & Octoprint SD: [image: 20200505_121630] https://user-images.githubusercontent.com/46318626/81057838-7f18e800-8ecd-11ea-8cea-64c8453562ad.jpg

I did suspect an issue with the baud rate and I swith my raspberry 3B to a raspberry 4B and things get a bit better but still not as good as the SD print. Raspberry 4 vs Raspberry 3: [image: 20200505_124726] https://user-images.githubusercontent.com/46318626/81059292-7a096800-8ed0-11ea-878d-a9f17bbfab8d.jpg

Raspberry 4 vs SD: [image: 20200505_124651] https://user-images.githubusercontent.com/46318626/81059328-8ab9de00-8ed0-11ea-82bc-c29dfec7a8eb.jpg

I am printing form the SD with same filament to verify. If there is still a reference, I will try to change the firmware for 115000 bauds.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/InsanityAutomation/Marlin/issues/134#issuecomment-623989340, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALGA64RKGOSIMSXCOW5HTOLRP7W5VANCNFSM4MYT5SVA .

chunter1 commented 4 years ago

Did you compile the firmware yourself and did you use the provided example config files for the CR10S Pro?

beny18241 commented 4 years ago

same here, using provided hex file.

FrancescoTalotta commented 4 years ago

I compiled the firmware with platformio, using the provided configuration for the CR10S Pro

chunter1 commented 4 years ago

Comparing the default "Configuration_adv" to the cr10spro specific one, i see the following difference in the serial buffer config:

DEFAULT: // The ASCII buffer for serial input ...

if ENABLED(MachineCR10Orig) //melzi has more ram than a 2560

define BUFSIZE 16

elif ANY(SKR13, SKR14, SKR14Turbo, SKRPRO11)

define BUFSIZE 8

else

define BUFSIZE 4

endif

CR10SPRO:

define BUFSIZE 4

So the default configures the buffer size to 16 and the cr10spro specific one to 4?

daredevil-ferreira commented 4 years ago

I did use the hex provided. The yellow filament turns ok when printed from the SD.

I just had a look to the 1.60.3 firmware provided by creality, the BUFSIZE is indeed set to 4.

As I understand the source code firmware, for the CR10sPro:

chunter1 commented 4 years ago
  • The configuration_adv.h in this case set the BUFSIZE to 4 also (we are not in the MachineCR10Orig nor SKR13, 14, 14Turbo, PRO11)

We are in the "MachineCR10Orig" since it is defined along the CR10S Pro Version.

daredevil-ferreira commented 4 years ago

In my configuration.h I have //#define MachineCR10Orig // Forces Melzi board

Later in the file

if ANY(MachineEnder2, MachineEnder3, MachineEnder5, MachineCR10, MachineMini) && NONE(Melzi_To_SBoardUpgrade, SKR13, SKRPRO11)

define MachineCR10Orig

endif

but all: //#define MachineEnder2 //#define MachineEnder3 //#define MachineEnder5 //#define MachineMini //#define MachineCR10

are also in comments.

I see that

if ENABLED(MachineCR10SPro)

define MachineCR10Std

But no MachineCR10Orig

But I will enforce BUFSIZE to 4 on my next trial

daredevil-ferreira commented 4 years ago

The firmware change at baud rate set to 115000 did not change quality. I am testing with BUFSIZE set to 4

chunter1 commented 4 years ago

I see that

if ENABLED(MachineCR10SPro)

define MachineCR10Std

But no MachineCR10Orig

Sorry, you are right. I confused MachineCR10Orig with MachineCR10Std.

daredevil-ferreira commented 4 years ago

No difference in quality of the print despite changing baud rate and BUFSIZE

I have no clue for any new testing.

daredevil-ferreira commented 4 years ago

Many several tries, last : BUFSIZE = 8 , BLOCK_BUFFER_SIZE = 16, baud rate back to 250000

Nothing works. I am going back to test the 1.70 version

chunter1 commented 4 years ago

Anyone found the reason or a fix for that issue yet? I just printed some stringing towers and there is definitive a huge problem when it comes to printing circles (=a lot of short motions in short short time). Went back to 1.1.9_B7 which works as expected.

daredevil-ferreira commented 4 years ago

No, and I found the same issue with the version 1.70. I was not able to find the first version of the firmware provided by Tiny Machines (I think they called it B6) to verify if the USB issue with circles was there. I did print a lot of items though with this first Tiny Machines firmware and I had never noticed this issue.

chunter1 commented 4 years ago

Since the Marlin 2 Version, the coordinates of the extruder are shown in realtime on the display. Could it be that the permanent serial communication to the display to update the values introduces a bottleneck?

daredevil-ferreira commented 4 years ago

where can I get the 1.1.9_B7 files ?

daredevil-ferreira commented 4 years ago

I will try to comment the SERIAL_ECHOPAIR_P(X_LBL, lpos.x, SP_Y_LBL, lpos.y, SP_Z_LBL, lpos.z, SP_E_LBL, lpos.e); in the report_logical_position of the motion.cpp file to see if this has an impact.

riri19 commented 4 years ago

Good idea

Le dim. 24 mai 2020 à 10:26 AM, daredevil-ferreira notifications@github.com a écrit :

I will try to comment the SERIAL_ECHOPAIR_P(X_LBL, lpos.x, SP_Y_LBL, lpos.y, SP_Z_LBL, lpos.z, SP_E_LBL, lpos.e); in the report_logical_position of the motion.cpp file to see if this has an impact.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/InsanityAutomation/Marlin/issues/134#issuecomment-633197389, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALGA64VGO7CEN54LCIUZ5W3RTDK3XANCNFSM4MYT5SVA .

chunter1 commented 4 years ago

where can I get the 1.1.9_B7 files ?

It's under the "Creality_DWINTest" branch: https://github.com/InsanityAutomation/Marlin/tree/Creality_DWINTest

chunter1 commented 4 years ago

I will try to comment the SERIAL_ECHOPAIR_P(X_LBL, lpos.x, SP_Y_LBL, lpos.y, SP_Z_LBL, lpos.z, SP_E_LBL, lpos.e); in the report_logical_position of the motion.cpp file to see if this has an impact.

Did you try it?

daredevil-ferreira commented 4 years ago

Hi not yet, I have a large print still on going. I will try this weekend.

Regards

Le mar. 26 mai 2020 à 17:24, chunter1 notifications@github.com a écrit :

I will try to comment the SERIAL_ECHOPAIR_P(X_LBL, lpos.x, SP_Y_LBL, lpos.y, SP_Z_LBL, lpos.z, SP_E_LBL, lpos.e); in the report_logical_position of the motion.cpp file to see if this has an impact.

Did you try it?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/InsanityAutomation/Marlin/issues/134#issuecomment-634094776, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALBMIIT4SQCK5LJ2JH6Z62LRTPNKPANCNFSM4MYT5SVA .

chunter1 commented 4 years ago

Unfortunately commenting out "SERIAL_ECHOPAIR_P(..." does not make a difference.

chunter1 commented 4 years ago

I just adapted the latest original Marlin version "bugfix-2.0.x" for the CR10S-Pro without display support. Hard to say, but i think the print is a bit better - however, i suspect the issue generally started with the transition to Marlin 2.0. Can it be that the focus on more performant 32-bit platforms leaves the 8-bit ones running into trouble?

mihaiserban commented 4 years ago

Same issue for me. upgraded a stock ender 5 plus to DW 6.2. I got these artifacts while printing round objects. I need to slow it down significantly to avoid them. 4qgqxjdr3b151

I noticed if i speed it up, there's a weird sound coming from the motors. Posted a video on reddit. https://www.reddit.com/r/ender5plus/comments/gnw43r/tinymachines_firmware_62_stock_ender_5_plus_weird/

daredevil-ferreira commented 4 years ago

I have the same issue with the 1.1.9 B7.

I've asked tiny machine where could I get the initial firmware (B6) I had.

jxlarrea commented 4 years ago

Confirmed. I'm also getting weird moire-like artifacts when printing round objects using the TinyMachines firmware on my Ender 5 Plus. Printing from OctoPrint.

wrx0EYJ

chunter1 commented 4 years ago

Hello. I noticed the same. Flashed the firmware 5 days ago and I have a lot of bubbles in the curved surfaces. I have a CR10s Pro with Octoprint. Do you know a commit that I can checkout in order to get rid of this problem?

image

Looks like the bandwidth issue of the UART communication as explained well in this video: https://www.youtube.com/watch?v=Hvw3DrVAeTA

mihaiserban commented 4 years ago

Kersey Fabrications has been running in similar quality issues with Marlin 2.0 - https://www.youtube.com/watch?v=smQ3JX_P0vE I'll try to disable //#define JD_USE_LOOKUP_TABLE as suggested in this Marlin 2.0 bug

chunter1 commented 4 years ago

Kersey Fabrications has been running in similar quality issues with Marlin 2.0 - https://www.youtube.com/watch?v=smQ3JX_P0vE I'll try to disable //#define JD_USE_LOOKUP_TABLE as suggested in this Marlin 2.0 bug

Isn't this only relevant when junction deviation is enabled which i don't have activated but am using classic jerk instead.

#if HAS_JUNCTION_DEVIATION
  // Enable this option for perfect accuracy but maximum
  // computation. Should be fine on ARM processors.
  //#define JD_USE_MATH_ACOS

  // Disable this option to save 120 bytes of PROGMEM,
  // but incur increased computation and a reduction
  // in accuracy.
  #define JD_USE_LOOKUP_TABLE
#endif
mihaiserban commented 4 years ago

Isn't this only relevant when junction deviation is enabled which i don't have activated but am using classic jerk instead.

indeed.. JUNCTION_DEVIATION is disabled by default in TM firmware right?

chunter1 commented 4 years ago

Isn't this only relevant when junction deviation is enabled which i don't have activated but am using classic jerk instead.

indeed.. JUNCTION_DEVIATION is disabled by default in TM firmware right?

Yes, disabled by default.

daredevil-ferreira commented 4 years ago

Anyway how this parameter can have a distinct impact regarding if the print is sent via Octoprint or the SD card ?

chunter1 commented 4 years ago

Has anyone got the Marlin 1.1.6 Version of Tiny Machines? As mentioned in the video below, the quality should be fine with it. https://www.youtube.com/watch?v=smQ3JX_P0vE

chunter1 commented 4 years ago

I'm switching to Klipper firmware since i wanna get rid of that annoying Creality DWIN display solution on my CR10S Pro anyway. [Update:] Just a quick update on Klipper running on my CR10S Pro. The documentation of Klipper is really good and installation is easy. Print of circles are running smooth and fast compared to Marlin. Besides the flawless print quality, the "on-the-fly" editable configuration is such an enhancement over Marlin. Further i got a 4,3" IPS display with capacitive touch to replace the really stupid one of Creality. So far, i can absolutely recommend the switch to Klipper.

daredevil-ferreira commented 4 years ago

What config did you use in Klipper ? I thought that Klipper was not running on the CR10s Pro (https://github.com/KevinOConnor/klipper/issues/1113)

Regards

chunter1 commented 4 years ago

What config did you use in Klipper ? I thought that Klipper was not running on the CR10s Pro

That issue is already fixed. I have setup my own Klipper config by using the example config in the wiki.

chunter1 commented 4 years ago

That's awesome!... With OctoKlipper you get performance graphs of your Klipper system :)

OctoKlipper

daredevil-ferreira commented 4 years ago

I am using Klipper and I am also very pleased. Bed Mesh leveling works even better than with marlin

Hoffi-D3 commented 4 years ago

What config did you use in Klipper ?

I thought that Klipper was not running on the CR10s Pro

That issue is already fixed.

I have setup my own Klipper config by using the example config in the wiki.

Hi

I have the same issue.

Could you please share your config with me? I want to give klipper a try. Thank you so much

chunter1 commented 4 years ago

Here is my Klipper config for the CR10S Pro so far. Please note, that I modded the TMC2208 on the Creality motherboard to support UART as described here: https://www.3dprintingforum.org/creality3d/126203-tmc2208-uart-on-cr10s-pro-for-linear-advance-guide#post126203 That way, i can use pressure advance (linear advance) on my CR10S Pro. :-) Further, for bed leveling, i am using a BLTouch. I highly recommend looking at the example config files Klipper provides in it's documentation.

[stepper_x] step_pin: ar54 dir_pin: ar55 enable_pin: !ar38 step_distance: .0125 endstop_pin: ^ar3 position_endstop: 0 position_max: 310 homing_speed: 80

[stepper_y] step_pin: ar60 dir_pin: !ar61 enable_pin: !ar56 step_distance: .0125 endstop_pin: ^ar14 position_endstop: 0 position_max: 310 homing_speed: 80

[stepper_z] step_pin: ar46 dir_pin: !ar48 enable_pin: !ar62 step_distance: .0025 endstop_pin: probe: z_virtual_endstop position_min: -5.0 position_max: 400

[extruder] step_pin: ar26 dir_pin: ar28 enable_pin: !ar24 step_distance: .00717641 nozzle_diameter: 0.400 filament_diameter: 1.750 heater_pin: ar10 sensor_type: EPCOS 100K B57560G104F sensor_pin: analog13 min_temp: 0 max_temp: 250 min_extrude_temp: 170 max_extrude_only_distance: 600 pressure_advance: 0.5 max_extrude_only_velocity: 200 max_extrude_only_accel: 3000

[tmc2208 extruder] uart_pin: ar33 microsteps: 16 interpolate: True run_current: 0.730 sense_resistor: 0.150

[heater_bed] heater_pin: ar8 sensor_type: ATC Semitec 104GT-2 sensor_pin: analog14 min_temp: 0 max_temp: 110

[fan] pin: ar9

[mcu] serial: /dev/serial/by-id/usb-FTDI_FT232R_USB_UART_AL05IRGY-if00-port0 pin_map: arduino

[printer] kinematics: cartesian max_velocity: 300 max_accel: 1500 max_z_velocity: 10 max_z_accel: 50 square_corner_velocity: 5

[safe_z_home] home_xy_position: 150,150 # Change coordinates to the center of your print bed speed: 50 z_hop: 10 # Move up 10mm z_hop_speed: 5

[force_move] enable_force_move: True

[bltouch] sensor_pin: ^ar18 control_pin: ar11 set_output_mode: 5V x_offset: -33.0 y_offset: -2.0 speed: 10

[bed_mesh] speed: 200 mesh_min: 50,50 mesh_max: 250,250 probe_count: 4,4

[filament_switch_sensor my_filemant_sensor] pause_on_runout: True pause_delay: 0.5 switch_pin: ar2

*# <---------------------- SAVE_CONFIG ---------------------->

*# DO NOT EDIT THIS BLOCK OR BELOW. The contents are auto-generated.

*

*# [bltouch]

*# z_offset = 2.140

*

*# [bed_mesh default]

*# version = 1

*# points =

*# 0.027500, -0.010000, 0.027500, 0.177500

*# 0.050000, -0.015000, 0.022500, 0.140000

*# -0.032500, -0.057500, -0.017500, 0.080000

*# 0.015000, -0.010000, 0.045000, 0.180000

*# x_count = 4

*# y_count = 4

*# mesh_x_pps = 2

*# mesh_y_pps = 2

*# algo = lagrange

*# tension = 0.2

*# min_x = 50.0000000001

*# max_x = 249.975

*# min_y = 50.0000000001

*# max_y = 249.975

*

*# [extruder]

*# control = pid

*# pid_kp = 26.636

*# pid_ki = 1.296

*# pid_kd = 136.840

*

*# [heater_bed]

*# control = pid

*# pid_kp = 67.769

*# pid_ki = 0.881

*# pid_kd = 1303.702

chunter1 commented 4 years ago

Bed Mesh leveling works even better than with marlin

I noticed that aswell. :)

Another thing i like is that there is no crypic EEPROM storage used. Instead, the values are appended to the config-file thus you can always have a look at the values stored.

Hoffi-D3 commented 4 years ago

Here is my Klipper config for the CR10S Pro so far.

Please note, that I modded the TMC2209 on the Creality motherboard to support UART as described here: https://www.3dprintingforum.org/creality3d/126203-tmc2208-uart-on-cr10s-pro-for-linear-advance-guide#post126203

That way, i can use pressure advance (linear advance) on my CR10S Pro. :-)

Further, for bed leveling, i am using a BLTouch.

I highly recommend looking at the example config files Klipper provides in it's documentation.

[stepper_x]

step_pin: ar54

dir_pin: ar55

enable_pin: !ar38

step_distance: .0125

endstop_pin: ^ar3

position_endstop: 0

position_max: 310

homing_speed: 80

[stepper_y]

step_pin: ar60

dir_pin: !ar61

enable_pin: !ar56

step_distance: .0125

endstop_pin: ^ar14

position_endstop: 0

position_max: 310

homing_speed: 80

[stepper_z]

step_pin: ar46

dir_pin: !ar48

enable_pin: !ar62

step_distance: .0025

endstop_pin: probe: z_virtual_endstop

position_min: -5.0

position_max: 400

[extruder]

step_pin: ar26

dir_pin: ar28

enable_pin: !ar24

step_distance: .00717641

nozzle_diameter: 0.400

filament_diameter: 1.750

heater_pin: ar10

sensor_type: EPCOS 100K B57560G104F

sensor_pin: analog13

min_temp: 0

max_temp: 250

min_extrude_temp: 170

max_extrude_only_distance: 600

pressure_advance: 0.5

max_extrude_only_velocity: 200

max_extrude_only_accel: 3000

[tmc2208 extruder]

uart_pin: ar33

microsteps: 16

interpolate: True

run_current: 0.730

sense_resistor: 0.150

[heater_bed]

heater_pin: ar8

sensor_type: ATC Semitec 104GT-2

sensor_pin: analog14

min_temp: 0

max_temp: 110

[fan]

pin: ar9

[mcu]

serial: /dev/serial/by-id/usb-FTDI_FT232R_USB_UART_AL05IRGY-if00-port0

pin_map: arduino

[printer]

kinematics: cartesian

max_velocity: 300

max_accel: 1500

max_z_velocity: 10

max_z_accel: 50

square_corner_velocity: 5

[safe_z_home]

home_xy_position: 150,150 # Change coordinates to the center of your print bed

speed: 50

z_hop: 10 # Move up 10mm

z_hop_speed: 5

[force_move]

enable_force_move: True

[bltouch]

sensor_pin: ^ar18

control_pin: ar11

set_output_mode: 5V

x_offset: -33.0

y_offset: -2.0

speed: 10

[bed_mesh]

speed: 200

mesh_min: 50,50

mesh_max: 250,250

probe_count: 4,4

[filament_switch_sensor my_filemant_sensor]

pause_on_runout: True

pause_delay: 0.5

switch_pin: ar2

*# <---------------------- SAVE_CONFIG ---------------------->

*# DO NOT EDIT THIS BLOCK OR BELOW. The contents are auto-generated.

*

*# [bltouch]

*# z_offset = 2.140

*

*# [bed_mesh default]

*# version = 1

*# points =

*# 0.027500, -0.010000, 0.027500, 0.177500

*# 0.050000, -0.015000, 0.022500, 0.140000

*# -0.032500, -0.057500, -0.017500, 0.080000

*# 0.015000, -0.010000, 0.045000, 0.180000

*# x_count = 4

*# y_count = 4

*# mesh_x_pps = 2

*# mesh_y_pps = 2

*# algo = lagrange

*# tension = 0.2

*# min_x = 50.0000000001

*# max_x = 249.975

*# min_y = 50.0000000001

*# max_y = 249.975

*

*# [extruder]

*# control = pid

*# pid_kp = 26.636

*# pid_ki = 1.296

*# pid_kd = 136.840

*

*# [heater_bed]

*# control = pid

*# pid_kp = 67.769

*# pid_ki = 0.881

*# pid_kd = 1303.702

Thank you! I will try :)

daredevil-ferreira commented 4 years ago

I added this macro to the printer.cfg file to be able to tune the Z offset.

================================== [gcode_macro G29] gcode: G1 Z10 F600

Sheet 1# SET_GCODE_OFFSET Z=-0.645

Sheet 2# SET_GCODE_OFFSET Z=-0.48

    SET_GCODE_OFFSET Z=-0.48
    BED_MESH_CALIBRATE
mihaiserban commented 4 years ago

I added this macro to the printer.cfg file to be able to tune the Z offset.

================================== [gcode_macro G29] gcode: G1 Z10 F600

Sheet 1# SET_GCODE_OFFSET Z=-0.645

Sheet 2# SET_GCODE_OFFSET Z=-0.48

SET_GCODE_OFFSET Z=-0.48 BED_MESH_CALIBRATE

Use the following to calibrate z offset in the terminal. Keep lowering the nozzle using TESTZ commands until you find the z offset. Then send an ACCEPT, and SAVE_CONFIG to save it in the printer.cfg file

PROBE_CALIBRATE TESTZ Z=-.1 ACCEPT SAVE_CONFIG