Running ODROID-C4/N2 with 3D printer

hominoid · Post by **hominoid** » Sun May 09, 2021 5:38 am

I’m getting really close to having everything running even though I’m still waiting on the replacement mosfets. I went ahead and started testing and debugging the other circuits. After enabling the klipper_mcu service and setting the address of the TMC2209’s correctly, all four stepper drivers can be configured through the UART and are reporting no errors in the klipper log file. When the heated bed thermistor was tested it did not report correctly at first. After some trouble shooting, I found the gpio pin assignments in the printer.cfg swapped between the heated bed and hot end. Easy enough to fix and both the heated bed and hot end thermistors are now also verified to work. As soon as mosfets arrive and can be installed, the heated bed, hot-end and fan outputs can be energized to complete the initial hardware startup tests.

At this point the klipper daemon seems to be working and seeing the proper devices but I have not connected stepper motors because I haven’t got the klipperscreen add on working in order to control them. It hasn’t worked from the first boot. The HDMI output shows the klipper splash screen and says it’s initializing but never completes. Stopping the klipperscreen service returns the HDMI output to a console login. I have not been able to access the moonraker api yet either using http://<printer_ip>:7125/printer/info per the moonraker setup instructions. I believe both problems are related. The moonraker api never connects and the klipperscreen.log and daemon status both show the following websocket connection refused error 111:

Code: Select all

hominoid@odroid-c4:/etc/default$ sudo service klipperscreen status
● klipperscreen.service - KlipperScreen
     Loaded: loaded (/lib/systemd/system/klipperscreen.service; enabled; vendor preset: enabled)
     Active: active (running) since Sat 2021-05-08 14:57:39 EDT; 24s ago
   Main PID: 2574 (xinit)
      Tasks: 28 (limit: 3463)
     Memory: 58.8M
     CGroup: /system.slice/klipperscreen.service
             ├─2574 /usr/bin/xinit /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/d>
             ├─2575 /usr/lib/xorg/Xorg :0
             └─2603 /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/kl>

May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,440 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,442 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,931 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,932 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/a>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,943 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,945 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,432 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,432 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/a>
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,442 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,444 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed

It appears to be a networking issue but I got concerned about the configuration when reviewing the python code because, I came across an incorrect hard-coded reference for the websocket api key which is currently only located in /root on this os image.

/usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlippyWebsocket.py

Code: Select all

#f = open("/home/pi/.moonraker_api_key", "r")
api_key = "" #f.readline()
#f.close()

I tried different possibilities(/root and ~/) and also added 127.0.0.1 to /etc/default/moonraker.conf under trusted users, per moonraker install instructions, but neither change helped. I reviewed the complete install and setup process of klipper, moonraker and mainsail per the developers documentation. At this point my next step might be to start with a new minimal image and try to build it all out manually but before I do...

@tobetter, do you have any suggestions? All the configuration files look correct. Is it possible for you to try a fresh copy of your os image and see if the klipperscreen service initializes and displays an interface? I even tried a vu7 and it looses video signal all together, like no display at all after the klipperscreen service starts. The printer.cfg file below has two corrections that should be in /usr/share/klipper/config/generic-odroidc4.cfg The heater_pin gpio# swap between the heated bed and extruder and the Odroid-N2/N2plus reference in the top comments changed to Odroid-C4.

config_files.zip: (2.68 KiB) Downloaded 52 times

/etc/klippy_config/printer.cfg
/etc/default/moonraker.conf
/usr/lib/python3.8/dist-packages/klipperscreen/ks_include/KlipperScreen.conf

tobetter · Post by **tobetter** » Sun May 09, 2021 11:37 am

hominoid wrote: ↑
Sun May 09, 2021 5:38 am
I’m getting really close to having everything running even though I’m still waiting on the replacement mosfets. I went ahead and started testing and debugging the other circuits. After enabling the klipper_mcu service and setting the address of the TMC2209’s correctly, all four stepper drivers can be configured through the UART and are reporting no errors in the klipper log file. When the heated bed thermistor was tested it did not report correctly at first. After some trouble shooting, I found the gpio pin assignments in the printer.cfg swapped between the heated bed and hot end. Easy enough to fix and both the heated bed and hot end thermistors are now also verified to work. As soon as mosfets arrive and can be installed, the heated bed, hot-end and fan outputs can be energized to complete the initial hardware startup tests.

Great...have you checked LED for fans and heaters if they change whenever the state is changed?
Also please note that you need to connect or wire the power source for the fans through the 2P connector at left-top, this connector is to supply 12V or 24V dependes on your fans.

At this point the klipper daemon seems to be working and seeing the proper devices but I have not connected stepper motors because I haven’t got the klipperscreen add on working in order to control them. It hasn’t worked from the first boot. The HDMI output shows the klipper splash screen and says it’s initializing but never completes. Stopping the klipperscreen service returns the HDMI output to a console login. I have not been able to access the moonraker api yet either using http://<printer_ip>:7125/printer/info per the moonraker setup instructions. I believe both problems are related. The moonraker api never connects and the klipperscreen.log and daemon status both show the following websocket connection refused error 111:
Code: Select all
hominoid@odroid-c4:/etc/default$ sudo service klipperscreen status
● klipperscreen.service - KlipperScreen
     Loaded: loaded (/lib/systemd/system/klipperscreen.service; enabled; vendor preset: enabled)
     Active: active (running) since Sat 2021-05-08 14:57:39 EDT; 24s ago
   Main PID: 2574 (xinit)
      Tasks: 28 (limit: 3463)
     Memory: 58.8M
     CGroup: /system.slice/klipperscreen.service
             ├─2574 /usr/bin/xinit /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/d>
             ├─2575 /usr/lib/xorg/Xorg :0
             └─2603 /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/kl>

May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,440 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,442 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,931 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,932 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/a>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,943 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:03 odroid-c4 xinit[2603]: 2021-05-08 14:58:03,945 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,432 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,432 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/a>
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,442 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connect>
May 08 14:58:04 odroid-c4 xinit[2603]: 2021-05-08 14:58:04,444 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
It appears to be a networking issue but I got concerned about the configuration when reviewing the python code because, I came across an incorrect hard-coded reference for the websocket api key which is currently only located in /root on this os image.
/usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlippyWebsocket.py
Code: Select all
#f = open("/home/pi/.moonraker_api_key", "r")
api_key = "" #f.readline()
#f.close()
I tried different possibilities(/root and ~/) and also added 127.0.0.1 to /etc/default/moonraker.conf under trusted users, per moonraker install instructions, but neither change helped. I reviewed the complete install and setup process of klipper, moonraker and mainsail per the developers documentation. At this point my next step might be to start with a new minimal image and try to build it all out manually but before I do...

Are you able to access Klipper using a web browser from your PC?
If KlipperScreen is not able to access, edit line number 20 in the file /etc/default/moonraker.con. I believe this will allow KlipperScreen to run after rebooting.

Code: Select all

 19 [authorization]
 20 enabled: False
 21 trusted_clients:

@tobetter, do you have any suggestions? All the configuration files look correct. Is it possible for you to try a fresh copy of your os image and see if the klipperscreen service initializes and displays an interface? I even tried a vu7 and it looses video signal all together, like no display at all after the klipperscreen service starts. The printer.cfg file below has two corrections that should be in /usr/share/klipper/config/generic-odroidc4.cfg The heater_pin gpio# swap between the heated bed and extruder and the Odroid-N2/N2plus reference in the top comments changed to Odroid-C4.
config_files.zip
/etc/klippy_config/printer.cfg
/etc/default/moonraker.conf
/usr/lib/python3.8/dist-packages/klipperscreen/ks_include/KlipperScreen.conf

Let me check the files and since I do not have Vu7 at the moment, I can test it later.

hominoid · Post by **hominoid** » Mon May 10, 2021 1:10 am

I don't have access to any klipper interface at the moment so I can't operate the stepper motors or turn on the bed or hot-end to check the leds. Disabling the authorization in moonraker.conf doesn't fix my interface problem. If I try accessing the printer from another browser on a different machine, there seems to be some recognition by the web server. The tab is titled Mainsail and I see two buttons before another dialog comes up indicating that it's connecting to the printer but never connects, as if moonraker is not responding.

if I check to see if port 7125 is open while all the services are running I don't see it.

Code: Select all

hominoid@odroid-c4:~$ netstat -lntu
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State      
tcp        0      0 127.0.0.53:53           0.0.0.0:*               LISTEN     
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN     
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN     
tcp6       0      0 :::22                   :::*                    LISTEN     
udp        0      0 127.0.0.53:53           0.0.0.0:*

I cannot go any further until I can access an interface so I'll keep working on it.

tobetter · Post by **tobetter** » Mon May 10, 2021 1:19 am

hominoid wrote: ↑
Mon May 10, 2021 1:10 am
I don't have access to any klipper interface at the moment so I can't operate the stepper motors or turn on the bed or hot-end to check the leds. Disabling the authorization in moonraker.conf doesn't fix my interface problem. If I try accessing the printer from another browser on a different machine, there seems to be some recognition by the web server. The tab is titled Mainsail and I see two buttons before another dialog comes up indicating that it's connecting to the printer but never connects, as if moonraker is not responding.

if I check to see if port 7125 is open while all the services are running I don't see it.
Code: Select all
hominoid@odroid-c4:~$ netstat -lntu
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State      
tcp        0      0 127.0.0.53:53           0.0.0.0:*               LISTEN     
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN     
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN     
tcp6       0      0 :::22                   :::*                    LISTEN     
udp        0      0 127.0.0.53:53           0.0.0.0:*                   
I cannot go any further until I can access an interface so I'll keep working on it.

Can you check if klippy and moonraker processes are running?

Code: Select all

$ ps aux | grep moon
$ ps aux | grep klipp

Also can you share the logs, /var/log/moonraker.log and /tmp/klippy.log?

hominoid · Post by **hominoid** » Mon May 10, 2021 1:56 am

Code: Select all

hominoid@odroid-c4:/tmp$ ps aux | grep moon
hominoid    2372  0.0  0.0   5836   652 pts/2    S+   12:38   0:00 grep --color=auto moon
hominoid@odroid-c4:/tmp$ ps aux | grep kipp
hominoid    2434  0.0  0.0   5836   636 pts/2    S+   12:38   0:00 grep --color=auto kipp

Code: Select all

hominoid@odroid-c4:/tmp$ sudo service moonraker status
[sudo] password for hominoid: 
● moonraker.service - Starts Moonraker on startup
     Loaded: loaded (/lib/systemd/system/moonraker.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:36:24 EDT; 33ms ago
   Main PID: 2110 (python3)
      Tasks: 1 (limit: 3463)
     Memory: 1.9M
     CGroup: /system.slice/moonraker.service
             └─2110 /usr/bin/python3 /usr/lib/python3/dist-packages/moonraker/moonraker.py -l /var/log/moonraker.log -c /etc/default/moonraker.conf

May 09 12:36:24 odroid-c4 systemd[1]: Started Starts Moonraker on startup.
hominoid@odroid-c4:/tmp$ sudo service klipper status
● klipper.service - Klipper service
     Loaded: loaded (/lib/systemd/system/klipper.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:31:37 EDT; 5min ago
   Main PID: 1743 (python2)
      Tasks: 4 (limit: 3463)
     Memory: 33.3M
     CGroup: /system.slice/klipper.service
             └─1743 /usr/bin/python2 /usr/lib/pypy/dist-packages/klippy/klippy.py /etc/klippy_config/printer.cfg -l /tmp/klippy.log -a /tmp/klippy_uds

May 09 12:31:37 odroid-c4 systemd[1]: Started Klipper service.
hominoid@odroid-c4:/tmp$ sudo service klipperscreen status
● klipperscreen.service - KlipperScreen
     Loaded: loaded (/lib/systemd/system/klipperscreen.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:31:37 EDT; 5min ago
   Main PID: 1747 (xinit)
      Tasks: 28 (limit: 3463)
     Memory: 124.0M
     CGroup: /system.slice/klipperscreen.service
             ├─1747 /usr/bin/xinit /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.c>
             ├─1767 /usr/lib/xorg/Xorg :0
             └─1890 /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.conf

May 09 12:36:53 odroid-c4 xinit[1890]: 2021-05-09 12:36:53,931 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:53 odroid-c4 xinit[1890]: 2021-05-09 12:36:53,933 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,422 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,423 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/access/oneshot_token
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,436 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,437 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,923 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,924 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/access/oneshot_token
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,933 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,934 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed

Moonraker's log file has no entry after the May 3 initial os image boot and subsequent system update and upgrade.

log_files.zip: (14.92 KiB) Downloaded 49 times

tobetter · Post by **tobetter** » Mon May 10, 2021 2:17 am

hominoid wrote: ↑

Mon May 10, 2021 1:56 am

Code: Select all

hominoid@odroid-c4:/tmp$ ps aux | grep moon
hominoid    2372  0.0  0.0   5836   652 pts/2    S+   12:38   0:00 grep --color=auto moon
hominoid@odroid-c4:/tmp$ ps aux | grep kipp
hominoid    2434  0.0  0.0   5836   636 pts/2    S+   12:38   0:00 grep --color=auto kipp

Code: Select all

hominoid@odroid-c4:/tmp$ sudo service moonraker status
[sudo] password for hominoid: 
● moonraker.service - Starts Moonraker on startup
     Loaded: loaded (/lib/systemd/system/moonraker.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:36:24 EDT; 33ms ago
   Main PID: 2110 (python3)
      Tasks: 1 (limit: 3463)
     Memory: 1.9M
     CGroup: /system.slice/moonraker.service
             └─2110 /usr/bin/python3 /usr/lib/python3/dist-packages/moonraker/moonraker.py -l /var/log/moonraker.log -c /etc/default/moonraker.conf

May 09 12:36:24 odroid-c4 systemd[1]: Started Starts Moonraker on startup.
hominoid@odroid-c4:/tmp$ sudo service klipper status
● klipper.service - Klipper service
     Loaded: loaded (/lib/systemd/system/klipper.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:31:37 EDT; 5min ago
   Main PID: 1743 (python2)
      Tasks: 4 (limit: 3463)
     Memory: 33.3M
     CGroup: /system.slice/klipper.service
             └─1743 /usr/bin/python2 /usr/lib/pypy/dist-packages/klippy/klippy.py /etc/klippy_config/printer.cfg -l /tmp/klippy.log -a /tmp/klippy_uds

May 09 12:31:37 odroid-c4 systemd[1]: Started Klipper service.
hominoid@odroid-c4:/tmp$ sudo service klipperscreen status
● klipperscreen.service - KlipperScreen
     Loaded: loaded (/lib/systemd/system/klipperscreen.service; enabled; vendor preset: enabled)
     Active: active (running) since Sun 2021-05-09 12:31:37 EDT; 5min ago
   Main PID: 1747 (xinit)
      Tasks: 28 (limit: 3463)
     Memory: 124.0M
     CGroup: /system.slice/klipperscreen.service
             ├─1747 /usr/bin/xinit /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.c>
             ├─1767 /usr/lib/xorg/Xorg :0
             └─1890 /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.conf

May 09 12:36:53 odroid-c4 xinit[1890]: 2021-05-09 12:36:53,931 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:53 odroid-c4 xinit[1890]: 2021-05-09 12:36:53,933 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,422 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,423 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/access/oneshot_token
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,436 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,437 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,923 - KlipperScreen.KlipperWebsocket - DEBUG - Attempting to reconnect
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,924 - KlipperScreen.KlippyRest - DEBUG - Sending request to http://127.0.0.1:7125/access/oneshot_token
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,933 - KlipperScreen.KlipperWebsocket - DEBUG - Websocket error: [Errno 111] Connection refused
May 09 12:36:54 odroid-c4 xinit[1890]: 2021-05-09 12:36:54,934 - KlipperScreen.KlipperWebsocket - DEBUG - Connection already closed

Moonraker's log file has no entry after the May 3 initial os image boot and subsequent system update and upgrade.
log_files.zip

Sorry, my bad...please run ps aux | grep klipp to check if klippy services are running. I have two processes like this.

Code: Select all

$ ps aux | grep klipp
root        1753  0.3  0.0   2252  1392 ?        S    17:09   0:00 /usr/share/klipper/firmware/LINUX/klipper_mcu -r
root        1816  1.6  0.7 254220 28436 pts/0    Ssl+ 17:09   0:03 /usr/bin/python2 /usr/lib/pypy/dist-packages/klippy/klippy.py /etc/klippy_config/printer.cfg -l /tmp/klippy.log -a /tmp/klippy_uds

I also have moonraker service running.

Code: Select all

$ ps aux | grep moon
root        1817  1.9  0.7 311968 28644 ?        Ssl  17:09   0:04 /usr/bin/python3 /usr/lib/python3/dist-packages/moonraker/moonraker.py -l /var/log/moonraker.log -c /etc/default/moonraker.conf

It's strage why moonraker.log only have May/3 logs. Can you try remove the log file and run sudo systemctl restart moonraker? This must start new service and drop log files for today.

By they way, did you update the package after flashing the image? I've found that recent packages - Klipper/Moonraker - from my repositories are not compatible...I need some work. So please do not update the package with sudo apt update before Printer starts to move.

hominoid · Post by **hominoid** » Mon May 10, 2021 2:48 am

That's interesting, I have more. I did immediately do an update and upgrade on this image right after the first boot and I noticed at the time it upgraded klipper, moonraker, mainsail and a bunch of python modules. FYI, I had tried deleting the moonraker.log in the past and a new one was not created. Luckily I saved the original log. I'm going to burn a new image and start over without doing an update and upgrade and see how it goes.

Code: Select all

hominoid@odroid-c4:~$ ps aux | grep klipp
root        1750  9.2  0.7 256444 27320 ?        Ssl  13:35   0:01 /usr/bin/python2 /usr/lib/pypy/dist-packages/klippy/klippy.py /etc/klippy_config/printer.cfg -l /tmp/klippy.log -a /tmp/klippy_uds
root        1752  1.0  0.0   2448  1576 ?        S    13:35   0:00 /usr/share/klipper/firmware/LINUX/klipper_mcu -r
root        1759  0.0  0.0   3876   892 ?        Ss   13:35   0:00 /usr/bin/xinit /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.conf
root        1896 16.6  2.5 1181568 95028 ?       Sl   13:35   0:02 /usr/bin/python3 /usr/lib/python3.8/dist-packages/klipperscreen/screen.py --configfile /usr/lib/python3.8/dist-packages/klipperscreen/ks_includes/KlipperScreen.conf

hominoid · Post by **hominoid** » Mon May 10, 2021 5:23 am

Starting again and not doing an update and upgrade worked. My thermistors are still reversed, need to look at that closer so ignore what I posted earlier about them. Thanks for the help @tobetter!

: Screenshot_2021-05-09_16-18-03.png (88.59 KiB) Viewed 7925 times

tobetter · Post by **tobetter** » Mon May 10, 2021 5:38 am

hominoid wrote: ↑
Mon May 10, 2021 5:23 am
Starting again and not doing an update and upgrade worked. My thermistors are still reversed, need to look at that closer so ignore what I posted earlier about them. Thanks for the help @tobetter!
Screenshot_2021-05-09_16-18-03.png

Cool!!

Let me share my configuration for thermistors...

Code: Select all

[extruder]
dir_pin: gpiochip0/gpio69
enable_pin: !gpiochip0/gpio68
step_pin: gpiochip0/gpio72
step_distance: 0.00179231
nozzle_diameter: 0.400
filament_diameter: 1.750
heater_pin: gpiochip1/gpio9
sensor_type: EPCOS 100K B57560G104F
sensor_pin: analog0
min_temp: 0
max_temp: 250

[heater_bed]
heater_pin: gpiochip1/gpio7
sensor_type: ATC Semitec 104GT-2
sensor_pin: analog2
#control: watermark
min_temp: 0
max_temp: 85

tobetter · Post by **tobetter** » Sun May 23, 2021 3:14 pm

20210523 New image

New OS image to run Klipper/MoonRaker/Mainsail on ODROID-C4 has been uploaded.
Click this to download, please note that this image will operate only after /etc/klippy_config/printer.cfg for your printer.

What are upgraded:

Klipper/Moonraker/Mainsail/KlipperScreen versions are upgraded.

Thumnail support has been fixed

Wifi connection can set up with nmtui in the command line.

Note:

If you have already installed and running 3D printer with the previous image 20200214, do run sudo apt update && sudo apt upgrade.

Thumbnail display is only supported when gcode file contains its thumbnail data and has been tested with Cura only. The format .ufp can be generated after installing a plug Cura2MoonrakerPlugin.

This image has been built with ODROID-Stamper, a tool to create ODROID image, with this command. If you are interesting what is ODROID-TAMPER, please visit the link.

sudo ODROID_STAMPER_CHECKOUT=$PWD ./odroid-stamper mainsail

.

: Screenshot from 2021-05-23 11-42-28.png (143.37 KiB) Viewed 7829 times

rooted · Post by **rooted** » Mon May 24, 2021 12:02 am

Awesome, thanks for keeping the image updated. I look forward to using it but haven't taken the plunge yet.

tobetter · Post by **tobetter** » Mon May 24, 2021 9:50 am

rooted wrote: ↑
Mon May 24, 2021 12:02 am
Awesome, thanks for keeping the image updated. I look forward to using it but haven't taken the plunge yet.

I am looking forward to see your printer runs with Klipper.

hominoid · Post by **hominoid** » Fri May 28, 2021 6:26 am

Parts have arrived so I’m continuing to work on this project and have staged a test configuration to verify the operation of each subsystem. There is one issue left but before I ask about it I want to share what I found regarding the thermistors.

: Staging_s.jpg (204.49 KiB) Viewed 7740 times

Even though when I initially posted about this issue I cited and changed the wrong pin in the Klipper printer configuration file

, I can only conclude at this point that my initial premise was correct. As you can see in the upper left corner, the heated bed thermistor J6(THB) is on the left and the hot-end thermistor J5(TH0) is the next connector to the right. This concurs with the schematic and the silk screen on the PCB. Using the default Klipper configuration for the Odorid-C4, which defines the heated bed as using analog2 and the hot-end using analog0, the reporting in both Mainsail and KlipperScreen are reversed. If I swap the connectors or change printer.cfg as indicated below, all interfaces report correctly.

Code: Select all

[extruder]
dir_pin: gpiochip0/gpio69
enable_pin: !gpiochip0/gpio68
step_pin: gpiochip0/gpio72
rotation_distance: 4.5714  # (200fsprx16ms)/700spmm
microsteps: 16
nozzle_diameter: 0.400
filament_diameter: 1.750
heater_pin: gpiochip1/gpio9
sensor_type: NTC 100K beta 3950
sensor_pin: analog2
control: watermark
min_temp: 0
max_temp: 320

[heater_bed]
heater_pin: gpiochip1/gpio7
sensor_type: EPCOS 100K B57560G104F
sensor_pin: analog0
control: watermark
min_temp: 0
max_temp: 110

The only two possible causes that I can think of is either the silkscreen and schematic are wrong or your thermistors connections are reversed.

One other thing I wanted to point out for anyone looking for a little more space and some openings under a Odroid-C4 with a shield, is the raised connector used for the GPIO. It’s 11mm tall and provides a total of 13mm space between the C4 PCB and the bottom of the shield's PCB and has some nice openings along the bottom. It is made by 3M part number 8540-4500PL

: gpio_connector_s.jpg (120.29 KiB) Viewed 7740 times

The last electronic issue I’m working on right now is solving why none of the mosfet circuits are working. I have tested all of the thermistors, end-stops and steppers and all work great. The heated-bed, hot-end and fans are not working. Throwing a oscilloscope on the input to the mosfet shows no signal on any of them when activated. When I went to reference the schematic to troubleshoot further, I couldn’t find any reference to their control signals(E0,HB,FAN0,FAN1) connected to anything except connector J31 and J12, which appear to be nothing but test points. Checking at J31 and J12 I found no signals and no 5v on those connectors either. Likewise, there is no reference to these signals on J9, the 40 pin gpio connector.

@tobetter, can you help me out here, there is obviously information I don’t have regarding this and where these control signals come from. Has there been changes to the schematic beyond the one I have? Can you please shed some light on what I don’t know(...as it relates to this problem because otherwise it would take way too long).

tobetter · Post by **tobetter** » Fri May 28, 2021 7:52 am

hominoid wrote: ↑
Fri May 28, 2021 6:26 am
Parts have arrived so I’m continuing to work on this project and have staged a test configuration to verify the operation of each subsystem. There is one issue left but before I ask about it I want to share what I found regarding the thermistors.
Staging_s.jpg
Even though when I initially posted about this issue I cited and changed the wrong pin in the Klipper printer configuration file , I can only conclude at this point that my initial premise was correct. As you can see in the upper left corner, the heated bed thermistor J6(THB) is on the left and the hot-end thermistor J5(TH0) is the next connector to the right. This concurs with the schematic and the silk screen on the PCB. Using the default Klipper configuration for the Odorid-C4, which defines the heated bed as using analog2 and the hot-end using analog0, the reporting in both Mainsail and KlipperScreen are reversed. If I swap the connectors or change printer.cfg as indicated below, all interfaces report correctly.
Code: Select all
[extruder]
dir_pin: gpiochip0/gpio69
enable_pin: !gpiochip0/gpio68
step_pin: gpiochip0/gpio72
rotation_distance: 4.5714  # (200fsprx16ms)/700spmm
microsteps: 16
nozzle_diameter: 0.400
filament_diameter: 1.750
heater_pin: gpiochip1/gpio9
sensor_type: NTC 100K beta 3950
sensor_pin: analog2
control: watermark
min_temp: 0
max_temp: 320

[heater_bed]
heater_pin: gpiochip1/gpio7
sensor_type: EPCOS 100K B57560G104F
sensor_pin: analog0
control: watermark
min_temp: 0
max_temp: 110
The only two possible causes that I can think of is either the silkscreen and schematic are wrong or your thermistors connections are reversed.

Oh...sorry about this, I've checked my C4 and it also switched the connectors. As long as Kipper can see the thermistor, it's ok.

One other thing I wanted to point out for anyone looking for a little more space and some openings under a Odroid-C4 with a shield, is the raised connector used for the GPIO. It’s 11mm tall and provides a total of 13mm space between the C4 PCB and the bottom of the shield's PCB and has some nice openings along the bottom. It is made by 3M part number 8540-4500PL
gpio_connector_s.jpg

Thank you, this is nice. I was looking for a part 1-2mm height more... It looks pretty good.

The last electronic issue I’m working on right now is solving why none of the mosfet circuits are working. I have tested all of the thermistors, end-stops and steppers and all work great. The heated-bed, hot-end and fans are not working. Throwing a oscilloscope on the input to the mosfet shows no signal on any of them when activated. When I went to reference the schematic to troubleshoot further, I couldn’t find any reference to their control signals(E0,HB,FAN0,FAN1) connected to anything except connector J31 and J12, which appear to be nothing but test points. Checking at J31 and J12 I found no signals and no 5v on those connectors either. Likewise, there is no reference to these signals on J9, the 40 pin gpio connector.

@tobetter, can you help me out here, there is obviously information I don’t have regarding this and where these control signals come from. Has there been changes to the schematic beyond the one I have? Can you please shed some light on what I don’t know(...as it relates to this problem because otherwise it would take way too long).

Don't you have any signal on 7 pins at all when you test with the controller board? Will you test at the pin header, please?
Can you please let me know the silkscreen of the PCB version, it must be on the bottom on your ODROID-C4.

By the way, I can only see the "ORBIT" stepper motor which I want to purchase and try sometime. I wonder how it works well.

hominoid · Post by **hominoid** » Fri May 28, 2021 9:42 am

The revision of the PCB is the same as the gerbers you posted, 2020/12/13. I knew that having the thermistors swapped wasn't a problem with the incorrect odroid-c4 printer configuration. The distributed odroid-c4 printer configuration is wrong and needed to be noted so others don't spend time on it needlessly.

The Orbiter looks sweet for sure, can't wait to get it on a printer to try it out. If you get one, strongly consider the LDO Orbiter instead of one of the knock-offs. There are subtle differences, the LDO body is injection molded fiber reinforced nylon while the others are not, I remember one competitors was ABS. Plus it was LDO's design and I personally want to reward open source designers, not the copy cats which are all about the same price anyway.

Your problably thinking the hot-end is a e3d volcano, it's not, it's a new product from one of the original J-head designers and manufacturers from back in the early reprap days.

hominoid · Post by **hominoid** » Sat May 29, 2021 12:40 am

Hey @tobetter, I was thinking if the gerbers need to change then I might be able to rework my current pcb if I have access to an updated schematic. If that doesn’t work then an updated set of gerbers will be needed to order a new pcb. Regardless, I’m having fun so don’t worry yourself.

rooted · Post by **rooted** » Sat May 29, 2021 2:59 am

I look forward to hearing about your experience with the Orbiter extruder, it's been on my radar for a good while.

You said it was created by LDO but Robert Lörincz is the guy who engineered the orbiter, does he work for LDO perhaps? I know he collaborates with them and gets a small percentage, he chose to only receive a small amount to keep the extruder affordable. Awesome guy.

hominoid · Post by **hominoid** » Sat May 29, 2021 5:18 am

Looks like your right, I assumed that he was a principle in the company or worked for them. I got that impression because the Thingiverse page shows a picture that has with words LDO Motors Oribiter v1.5 but I see it also says in the text "This extruder version was designed in cooperation with LDO Motors and Vivedino, without their great support this product would not exist!", Thanks for the correction @rooted.

tobetter · Post by **tobetter** » Sat May 29, 2021 4:50 pm

hominoid wrote: ↑
Fri May 28, 2021 9:42 am
The revision of the PCB is the same as the gerbers you posted, 2020/12/13. I knew that having the thermistors swapped wasn't a problem with the incorrect odroid-c4 printer configuration. The distributed odroid-c4 printer configuration is wrong and needed to be noted so others don't spend time on it needlessly.

Oh, I meant the revision of ODROID-C4 itself. There is only one PCB for the 3D printer from me yet.

Assuming that your ODROID-C4 is engineering version before C4 had been released which revision must be 0.3, pinout on J7 is different from the mass production one. I want you to try to use these GPIO names in your printer.cfg for FETs.

Code: Select all

 
#1 -> GND
#2 -> GPIOAO_BIT10 (gpiochip1/gpio11)
#3 -> +5V
#4 -> GPIOAO_BIT9  (gpiochip1/gpio9)
#5 -> GPIOX_BIT10  (gpiochip0/gpio75)
#6 -> GPIOX_BIT11  (gpiochip0/gpio76)
#7 -> GPIOX_BIT9   (gpiochip0/gpio74)

Unfortunately, I don't have 0.3 version as of now...so I am not able test it...I will try to find one later. Please try and let me know which works and not, also you should not use SD card to run your ODROID-C4.

By the way, are you properly run Y steppers and Z-endstop with the current ODROID-C4? Both also has different pinout...

hominoid · Post by **hominoid** » Sun May 30, 2021 12:34 pm

I’m using a Rev 1.0 Odroid-C4 with an emmc so no problem there. I didn’t understand what you wanted me to do with the gpio names anyway because the syntax you used doesn’t make sense to me and looking at the klipper config reference didn’t help either. I tried some different syntax but klipper would never start and would complain about my changes. I moved on since I didn’t have a rev 0.3 odroid and I wasn’t making progress. I need to step back because I feel like I've missed something that I need to understand before I can move forward.

Where does E0,HB,FAN0 and FAN1 come onto the shield. The schematic doesn’t show any input for these signals. Using both a pcb and the gerbers I can visually see and physically ohm traces backwards from where E0 leaves the board, fed at the mosfet Q3 pin1 to J12,J31,R7 and R8 but nowhere else. I used an ohm meter at R7 to all 40 pins at J9 and found no connections. All 40 pins have signals assigned and none of them are E0,HB,FAN0 or FAN1. Is there some other connection or fundamental concept I missed? Does the Odroid-C4 J7 get connected somewhere? What am I missing here?

tobetter · Post by **tobetter** » Sun May 30, 2021 5:17 pm

hominoid wrote: ↑
Sun May 30, 2021 12:34 pm
Where does E0,HB,FAN0 and FAN1 come onto the shield. The schematic doesn’t show any input for these signals. Using both a pcb and the gerbers I can visually see and physically ohm traces backwards from where E0 leaves the board, fed at the mosfet Q3 pin1 to J12,J31,R7 and R8 but nowhere else. I used an ohm meter at R7 to all 40 pins at J9 and found no connections. All 40 pins have signals assigned and none of them are E0,HB,FAN0 or FAN1. Is there some other connection or fundamental concept I missed? Does the Odroid-C4 J7 get connected somewhere? What am I missing here?

Sorry, I think I misunderstood your previous post with photo...I've checked the photo of your board and can see J12 is missing, is the connector still missing or your board? This is from J7 of ODROID-C4, that's why the signal is missing. J12 is for ODROID-C4 and J31 is for ODROID-N2 so E0/HB/FAN0/FAN1 can be controlled through the pins. Can you please check it again?

: Staging_s.jpg (75.33 KiB) Viewed 7606 times

hominoid · Post by **hominoid** » Sun May 30, 2021 11:12 pm

I don’t have J12 or J31 populated so that is going to be the problem. It’s Sunday morning here so as soon as I finish my coffee I’ll install a connector. I’m confident that this is going to solve the problem, it’s the piece of the puzzle I missed. I looked at the pictures you posted many times to try and discern various information but there wasn’t a picture of the bottom of the board and the best one of the top didn’t appear to have them installed as far as I could tell. There is another connector that doesn’t appear to be used, J30. What is the purpose of this connector or is it just test points? Thanks again!

tobetter · Post by **tobetter** » Sun May 30, 2021 11:19 pm

hominoid wrote: ↑
Sun May 30, 2021 11:12 pm
I don’t have J12 or J31 populated so that is going to be the problem. It’s Sunday morning here so as soon as I finish my coffee I’ll install a connector. I’m confident that this is going to solve the problem, it’s the piece of the puzzle I missed. I looked at the pictures you posted many times to try and discern various information but there wasn’t a picture of the bottom of the board and the best one of the top didn’t appear to have them installed as far as I could tell. There is another connector that doesn’t appear to be used, J30. What is the purpose of this connector or is it just test points? Thanks again!

Sorry, for making a confusion...I had to upload more documents / pictures.
J30 is to connect SPI chip select pins when SPI type motor drivers are mounted rather than UART type. If your motor drivers are all UART types, like TMC2209, J30 can be ignored. Also, I believe you have properly set the address of your stepper drives with J21, J22, J23, J24 in UART mode.

hominoid · Post by **hominoid** » Sun May 30, 2021 11:29 pm

tobetter wrote: ↑
Sun May 30, 2021 11:19 pm
Sorry, for making a confusion...I had to upload more documents / pictures.
J30 is to connect SPI chip select pins when SPI type motor drivers are mounted rather than UART type. If your motor drivers are all UART types, like TMC2209, J30 can be ignored. Also, I believe you have properly set the address of your stepper drives with J21, J22, J23, J24 in UART mode.

It's not your fault, you didn't create any confusion, this is to be expected doing new and different undocumented things. I really appreciate you doing this project and I'm very grateful for your patience while helping with my questions. In reality I should have figured this out on my own earlier but I don't think I have ever used J7 on the C4 so it wasn't at the front of my mind while working on this project. I'll let you know latter how it went but as I said in my previous post, I'm confident everything will be fine now.

hominoid · Post by **hominoid** » Mon May 31, 2021 5:14 am

Good news, leds are flashing, fans are blowing, steppers are spinning and the hot-end and bed are getting hot. Everything is reporting and working as designed! I’ll start now to focus on getting all this gear on a frame so plastic can be spewing as well! There’s about to be another Odroid sbc driven 3d printer in the world.

tobetter · Post by **tobetter** » Mon May 31, 2021 9:34 am

hominoid wrote: ↑
Mon May 31, 2021 5:14 am
Good news, leds are flashing, fans are blowing, steppers are spinning and the hot-end and bed are getting hot. Everything is reporting and working as designed! I’ll start now to focus on getting all this gear on a frame so plastic can be spewing as well! There’s about to be another Odroid sbc driven 3d printer in the world.

Congratulations!! Please share us how your new printer design works...

hominoid · Post by **hominoid** » Sat Jun 05, 2021 11:33 pm

Here is another application processor based 3D printer control board that looks interesting. Things are heating up in this area of 3D printing.

rooted · Post by **rooted** » Sun Jun 06, 2021 4:25 am

hominoid wrote:Here is another application processor based 3D printer control board that looks interesting. Things are heating up in this area of 3D printing.

It's an interesting device but I still think external drivers are where it's at, most people aren't capable of replacing damaged surface mount components.

hominoid · Post by **hominoid** » Sun Jun 06, 2021 7:48 am

rooted wrote: ↑
Sun Jun 06, 2021 4:25 am
It's an interesting device but I still think external drivers are where it's at, most people aren't capable of replacing damaged surface mount components.

Being able to replace or upgrade the drivers is important for me too. I also like the idea of using an off-the-shelf SBC much better then a special purpose design like many are doing, easier to repair, upgrade and better SOC and memory specs.

rooted · Post by **rooted** » Sun Jun 06, 2021 9:42 am

hominoid wrote: ↑
Sun Jun 06, 2021 7:48 am

rooted wrote: ↑
Sun Jun 06, 2021 4:25 am
It's an interesting device but I still think external drivers are where it's at, most people aren't capable of replacing damaged surface mount components.
Being able to replace or upgrade the drivers is important for me too. I also like the idea of using an off-the-shelf SBC much better then a special purpose design like many are doing, easier to repair, upgrade and better SOC and memory specs.

I do appreciate the design and am glad you shared it, I follow CNX software but don't always visit everyday and missed it.

hominoid · Post by **hominoid** » Mon Jun 21, 2021 5:11 am

It’s been a couple of weeks since my last update and it’s still going to be a little longer until I have a printer assembled and running with the 3DPrinterShield, so it seems like a good time to bring everyone up to date on what’s happened so far. While considering what to do with the 3DPrinterShield, I initially though about building a new printer but I want to build a xy-core style and the 3DPrinterShield is not ready for that...yet. Since I have a xz-core printer I couldn’t justify building another so I decided to upgrade my existing printer instead.

Now that is no small task since it’s so old. Josef Prusa’s i3 design was introduced in May of 2012 and he made his first commercial MK0 version in 2015. I built my then, state of the art, pre-mk0 i3 in the winter of 2013 after research and planning thru the fall 2012. It is based on the original specifications that used 3mm filament, 5mm z-axis threaded rods and doesn’t include other changes and improvements that have taken place over the years. In order to run and test closer to the 3DPrinterShield’s capabilities, all three axis, the x-carriage, extruder and hot-end are being completely changed out to bring it up to date with the latest technology.

: i3_upgrades.jpg (202.95 KiB) Viewed 7189 times

I have been assembling and checking sub-assemblies as I go because I won’t be able to produce parts once the final assembly starts. Here’s a brief summary of the significant changes:

Changed from 3mm to 1.75mm filament.
Changed z-axis 5mm threaded rod to 8mm lead screws.
Upgraded z-axis motors mounts and rod holders.
Upgraded x-axis ends to support 8mm lead screws.
Upgraded x-axis idler pulley.
Upgraded to y-axis bearing holders.
Upgraded y-axis idler pulley.
Increased y-axis single llm8u bearing (24mm) side to a llml8u bearing (45mm).
New LDO Bearings on all axis.
Redesign of y-axis to change from 6mm to 9mm belt.
New custom x-carriage.
New Orbiter extruder and J-head Hi-flow hot-end mount.
Upgraded 6mm and 9mm belts to Gates 2GT Powergrip belts
All idlers and pulleys upgraded to Gates or LDO
All idlers installed on hardened and chromed steel dowels
New x and y-axis .09 degree, high temp (180c) LDO motors.
New z-axis LDO motors with integrated 8mm lead screw.
New single or isolated dual channel 60A MOSFET driver
New Re-enforce z-axis to y-axis frame connection.

So in the end, I did build a new printer but reused the frame, rod set, y-carriage and table. The most interesting of the changes is the new custom x-carriage and mount for the Orbiter extruder and J-head Hiflow hot-end, as well as the new y-axis 9mm belt system and increased bearing size. I’m hoping this will give me better acceleration and control for the y-axis; having the aluminum y-carriage, bed heater, table and printed part to move. I don’t think I’ll be able to reach @tobetter’s 150mm/s of his modified Ender-3, but I think I might be able to get close. Wheels are bound to be faster and lighter then bearings and rods but, I suspect the bearings will last longer and hold their accuracy better then a wheel in the long run.

: Orbiter-Mount.gif (1.64 MiB) Viewed 7189 times

To expedite the creation of the Orbiter mount, I started with the MK3 x-carriage(source, stl's) and mount, redesigning it to remove the Prusa builtin extruder and other unnecessary parts while adding hidden cable conduits and changes to accommodate the Orbiter, J-head, cooling and lighting system.

As I write, I’m printing what I hope are the final changes to the electronics enclosure(est. 9.5 hrs @ 40mm/s using 12 meters of 3mm filament for the case bottom), which should be the last printed parts needed. Yesterday I received the silicon insulated wire I ordered but unfortunately they sent the wrong gauge for one of the rolls. I’ll need to fix that and I’m still waiting on wire sheathing to arrive, before I can start final assembly.

rooted · Post by **rooted** » Mon Jun 21, 2021 2:10 pm

I don't think you will have any issue running up to around 150mm/s once you get your new build fully tuned. I really like those steppers with integrated leadscrews, one less connection you have to worry about.

tobetter · Post by **tobetter** » Wed Jun 23, 2021 4:11 am

@hominoid, I wonder what's the brand of your filament also wonder if there is a benefi to use wider timing belt. Your timing belt does not look like 6mm...

hominoid · Post by **hominoid** » Wed Jun 23, 2021 7:30 am

tobetter wrote: ↑
Wed Jun 23, 2021 4:11 am
@hominoid, I wonder what's the brand of your filament also wonder if there is a benefi to use wider timing belt. Your timing belt does not look like 6mm...

The belt in the photo above is the 9mm belt for the y-axis. I looked at upgrading the x-axis to 9mm but there was some interference so I decided to wait, it’s still 6mm. The filament is a no name brand flat black ABS. It’s the last black I found in 3mm, it's way out of round and the volumetrics change throughout the roll. The only thing available lately in 3mm is odd colors and glow in the dark, unless you want to pay the scalpers high prices. I’ll be happy to be on 1.75mm filament soon.

tobetter · Post by **tobetter** » Wed Jun 23, 2021 8:59 am

hominoid wrote: ↑
Wed Jun 23, 2021 7:30 am

tobetter wrote: ↑
Wed Jun 23, 2021 4:11 am
@hominoid, I wonder what's the brand of your filament also wonder if there is a benefi to use wider timing belt. Your timing belt does not look like 6mm...
The belt in the photo above is the 9mm belt for the y-axis. I looked at upgrading the x-axis to 9mm but there was some interference so I decided to wait, it’s still 6mm. The filament is a no name brand flat black ABS. It’s the last black I found in 3mm, it's way out of round and the volumetrics change throughout the roll. The only thing available lately in 3mm is odd colors and glow in the dark, unless you want to pay the scalpers high prices. I’ll be happy to be on 1.75mm filament soon.

Oh, thanks.I was curious what color is but does not look like a black. One thing a bit surprise is it's 3mm.

I am looking forward your new printer.

hominoid · Post by **hominoid** » Mon Jun 28, 2021 10:25 am

After receiving my last shipment and finishing everything on my list, the physical upgrade is underway. Working on the frame and mechanical upgrades first before moving on to the electronics and wiring. While also having a small print-fest this weekend to use up a couple of partial 3mm filament rolls, I pulled together all the printed parts and frame into a model to check for any clearance or alignment issues. Next post should be about a working Odroid SBC driven 3D printer or it’s issue(s).

: i3_frame_upgrade.gif (5.21 MiB) Viewed 6971 times

The above model for anyone that wants to take a closer look.

i3_frame_upgrade.zip: (3.32 MiB) Downloaded 25 times

rooted · Post by **rooted** » Mon Jun 28, 2021 12:09 pm

What is your center frame material? And the bed mount?

hominoid · Post by **hominoid** » Mon Jun 28, 2021 12:18 pm

rooted wrote: ↑
Mon Jun 28, 2021 12:09 pm
What is your center frame material? And the bed mount?

The frame and carriage are 6mm aluminum. I run a 4mm aluminum table on top of the bed heater as well.

rooted · Post by **rooted** » Mon Jun 28, 2021 3:38 pm

hominoid wrote:
rooted wrote: ↑
Mon Jun 28, 2021 12:09 pm
What is your center frame material? And the bed mount?
The frame and carriage are 6mm aluminum. I run a 4mm aluminum table on top of the bed heater as well.

Nice and sturdy I imagine

I really need to replace my heated bed, the only reason I haven't is it has neodymium magnets underneath and I don't want to lose that ability. I would have to find a custom 250x250 heater with cutouts for magnets unless I could somehow detach the heater from my current bed which seems unlikely without damaging it.

It came from the factory warped and I got a replacement but it's just as warped. I don't know why they most manufacturers go with 3mm instead of 4, the amount of beds they replace must nearly exceed the cost savings.

I know I could just use a magnetic sticker but I would much rather have the high temperature neodymium magnets.

hominoid · Post by **hominoid** » Mon Jun 28, 2021 11:09 pm

I have read that a lot of people are having problems with 3mm aluminum beds warping. I had trouble with the Borosilicate glass I started with because it wasn't very flat. I replaced it early on with the decked piece of 4mm aluminum I'm currently using and haven't had any problems since. I've been thinking about ways to incorporate a textured spring sheet into my configuration with or without using magnets, haven't thought of an easy and good way yet.

hominoid · Post by **hominoid** » Sun Jul 11, 2021 7:39 am

PRUSA i3 Odroid Edition
The i3 upgrades and electronic conversion to @tobetter’s 3DPrinter Shield design running on an Odroid-C4 is operational and works great. It’s a new printer in all aspects of performance but I haven’t had an opportunity yet to push it to it’s limit, to see just how fast it can go.

: i3_odroid.jpg (142.57 KiB) Viewed 6810 times

There’s a few things I need to do like more tuning and print a few spare parts to have on hand in case I break something. Since I can now print at much higher temps(300c+), I might even print some in nylon. There are also a few minor changes to do including additional cable management for the electronics enclosure and a new y-axis end-stop holder. The upgraded y-axis bearing holders raised my bed assembly by a few mm and it doesn’t align well with the end-stop.

The electronics are currently powered at 12v 30amps with two feeds of a 12awg pair, one for the MOSFET and one for the Odorid-C4 and shield. The MK2b heated bed is dual voltage and is setup for 12v from the MOSFET. The hi-flow hot-end can be setup for either 12v or 24v but is currently 12v directly from the shield.

: i3_elec.jpg (183.54 KiB) Viewed 6810 times

The frame upgrades, especially the new 9mm belt driven y-axis and the y-z axis braces, worked out well. I had to drill and tap 4 new holes in the frame for the braces but they provide a significant improvement in overall frame rigidity.

: i3_frame-support.jpg (96.54 KiB) Viewed 6810 times

There were several Klipper printer configuration changes made and I had a couple hardware issues to troubleshoot during the subsequent startup. The 7pin connection between the shield and the C4 isn’t quite long enough to engage 100% since I used a taller 40pin GPIO connector. I’ll have to look for a taller 7 pin connector long-term but, by bending the pins on the C4 away from the heatsink by about .25mm, a reliable connection could be established. The heated bed initially wouldn’t heat either, even though it did as part of the initial pre-build test; found a cold soldered surface mount 603 resistor on the shield.

: i3_head.jpg (148.72 KiB) Viewed 6810 times

: i3_odroid-2.jpg (115.39 KiB) Viewed 6810 times

This project has been a lot of fun and I look forward to the higher quality and faster prints these upgrades bring. I have to admit it scares me a little to see my printer move as fast as it does now with acceleration. I can’t wait to be able to really push it and, with all of the improvements and being an Odroid SBC driven 3D printer, makes this state-of-the-art once again too, for it’s class imo...serial# 2.

Thanks @tobetter for the 3DPrinter Shield and your work getting Klipper running on Odroid SBC!

tobetter · Post by **tobetter** » Sun Jul 11, 2021 12:10 pm

Thank you @hominoid, this looks really nice and beautiful. I am also curious how it moves / prints fast...
I also like the light at hotend, makes me to install light to my printer.

But I wonder why there are no cables for end-stops??

Great work and thank you for sharing the design.

rooted · Post by **rooted** » Sun Jul 11, 2021 7:15 pm

tobetter wrote: But I wonder why there are no cables for end-stops??

Great work and thank you for sharing the design.

Sensorless homing I imagine, or does your board not support it?

tobetter · Post by **tobetter** » Sun Jul 11, 2021 7:17 pm

rooted wrote: ↑
Sun Jul 11, 2021 7:15 pm

tobetter wrote: But I wonder why there are no cables for end-stops??

Great work and thank you for sharing the design.
Sensorless homing I imagine, or does your board not support it?

I've not wired the connection since I had no luck to make sensorless homing.

hominoid · Post by **hominoid** » Mon Jul 12, 2021 2:50 am

tobetter wrote: ↑
Sun Jul 11, 2021 7:17 pm

rooted wrote: ↑
Sun Jul 11, 2021 7:15 pm

tobetter wrote: But I wonder why there are no cables for end-stops??

Great work and thank you for sharing the design.
Sensorless homing I imagine, or does your board not support it?
I've not wired the connection since I had no luck to make sensorless homing.

Sorry for the slow response, my ISP has been down since last night...
All of the above. I was planning on using sensorless homing when these photo's were taken but when I went to configure it and found out that the diag_pin needed to be connected to a GPIO and it looked like it wasn't, I used the mechanical end-stops and rigged the y-axis holder to temporarily work. They just snap on to the rods but to make them permanent I need to add some more cable management for the case .

The light at the hot-end is a really nice feature. The one I'm using is 12v from an automobile lighting system and works very well. I used the same one on the original printer build which came as a 2 pack, so I'll already had a new one to use for this build.

rooted · Post by **rooted** » Mon Jul 12, 2021 7:44 pm

Do you use the light to record timelapse or just to better see the print?

I've got an LED strip to put on mine for general lighting but have yet to install it, I really like these newer style super dense COB strips and they just so happen to slide into the aluminum extrusion channels.

hominoid · Post by **hominoid** » Mon Jul 12, 2021 9:50 pm

rooted wrote: ↑
Mon Jul 12, 2021 7:44 pm
Do you use the light to record timelapse or just to better see the print?

I've got an LED strip to put on mine for general lighting but have yet to install it, I really like these newer style super dense COB strips and they just so happen to slide into the aluminum extrusion channels.

Just for light so I can see whats happening at the nozzle and to judge the quality of the print. I don't have video monitoring

hominoid · Post by **hominoid** » Mon Aug 02, 2021 11:17 pm

After buttoning up a few loose ends I started running 3DBenchy test prints in ABS at increasing speeds to probe were the performance envelope is for this machine and I have been pleasantly surprised. I get relatively decent prints at a multitude of speeds and there is room for improvement and the possibility to go faster. The creators of the benchmark print suggests print speeds of less then or equal to 50mm/s and 150mm/s travel speed which I'm using as a quality reference.

: 3dbenchy.jpg (192.02 KiB) Viewed 6546 times

I’m interested in maximizing print speeds for everyday use so all prints were sliced at typical default speed ratios used by slic3r and cura. Even then, the 3DBenchy prints in reality were limited to 500mm/s maximum speed due to it’s small size. At the current acceleration setting there isn’t enough room to accelerate to higher speeds so a different test model is needed to explore print speeds greater than 500mm/s. I have done a couple of short high speed runs, up to approximately 860mm/s while printing other larger parts, but not a complete print yet.

: 50vs500-2.jpg (94.24 KiB) Viewed 6546 times

At 500mm/s the printer doesn’t show any signs of being overly stressed, the motors are warm but not excessively hot. The Odroid-C4 is using between 25%-35% of a single core and runs cool even in the enclosure. I have not noticed any skipped steps or other unexplained problems. At print speeds of 300mm/s or greater more part cooling is definitely needed due to the short layer times and a nozzle temperature of 230c to get adequate flow. I believe inadequate cooling may be responsible for the occasional dislodged strands at those speeds. I have been considering varies ways to add more cooling, including a change over to air jets instead of a fan. Pressure advance is also being used but I’m not sure that it is set optimally right now.

: 50vs500-1.jpg (102.89 KiB) Viewed 6546 times

There is a definite increase in print resonance as speeds exceed 150mm/s so resonance compensation might help and needs to be explored moving forward. I don’t have either the 6mm or 9mm belt all that tight so tightening up the belts may help, as well as other possibilities. There is more tweaking to do.

This is a very short clip of the 500mm/s print, first layer at half speed. Unfortunately I don’t have a way to share the whole video.

i3-Odroid_Edition-short_1280x720-2.mp4.zip: (9.91 MiB) Downloaded 23 times

tobetter · Post by **tobetter** » Sat Aug 07, 2021 6:18 am

I've received the full video from @hominoid how his printer runs with ODROID-C4 and permitted to publish the video. This is very interesting that the movement of bed and head looks very stable than what I tried with Ender-3. The printing speed is 500 and acceleration is 3500 according to him.

Appreciate to @hominoid for the challenge and sharing his effort, very happy to see someone else tried this.

rooted · Post by **rooted** » Sat Aug 07, 2021 7:50 am

@hominoid It looks like you are printing the two inner walls slower and the outer wall fast? That's backwards to get a higher quality finish.

@tobetter I would love to do the same but unfortunately it's beyond my skills, I can't even get klipper to run on my stock board as you know :/