Re: Rpi Zero & Octopi working!!!!
Hello to all!
i've got my new rpi (after i burnt one) and regarding Linux i'm pretty much newbie... as a result, i'm stuck at adding "/dev/ttyAMA0". ... where to add this line? I guess because of this line is missing (somewhere) in octoprint i don't have any serial ports available...
Thanks!
This is an Octoprint setting.
Go to the wrench on the top. The first item is Serial Connection. There is a big text field for Additional Serial Ports. Add it there.
Re: Rpi Zero & Octopi working!!!!
@scott.w12 I still cant Mk3 work with octopi, I followed the instructions you give, keep getting "Unhandled communication error" for scanning baudrates. Error message: No more daudrates to test, and no suitable baudrate found.
Re: Rpi Zero & Octopi working!!!!
Andrew, many thanks! I managed to connect my RPi !
Now i have to do some tests. Among others i wonder - I can monitor/control printer via my phone, but can this setup send push notofication on my phone when print is finished or something goes wrong?
Also - it seems that g code monitoring on my PC is possible only if i choose to print through octoprint, while if i select print from SD i can only monitor temperatures...?
Re: Rpi Zero & Octopi working!!!!
I had previously been working on mounting a Pi Zero and camera to the left side Y axis pulley. I may go back to that idea, and wire the Pi into to Einsy, rather than directly. Might be cleaner than splicing together camera cables. Would also make the Pi much easier to access.
This was the route I took. It seemed much easier to gracefully maneuver a bundle of six 22awg hook-up wires from the Einsy to the front of the printer (rather than attempt the same with the RPi flat ribbon cable + coupler).
Thanks to scott.w for the detailed writeup. Basic connectivity is working fine for me, however printing via OctoPrint fails almost instantly just as others have described. Hopefully the PRUSA team is working on a firmware fix for this. The MK3 has been advertised as "OctoPrint Ready" since September.
Re: Rpi Zero & Octopi working!!!!
Did you install and try latest FW? Supposely they fixed that...?
Re: Rpi Zero & Octopi working!!!!
Some of the features that was making octoprint crash have been disabled in version 1.38 josef said
Hopefully they solve this quickly,,,,, Connecting the Pi and getting it to communicate with the printer was actually quite easy
Re: Rpi Zero & Octopi working!!!!
I installed the RC3 firmware yesterday, Octoprint is still having issues. On RC2 I was having lots of issues with prints failing to start printing. The printer would level the bed, and then halt. That seems to have cleared up under RC3, however now the filament sensor is acting weird, prompting to change filament multiple times mid print. It also was doing erratic movements, like it was checking end stops sporadically during the print, and eventually resulting in layer shifts.
I am also having issues when using Octoprint externally on a RPi 2. This has been frustrating. I used Octoprint for a year on my old Maker Select, it worked great. When this MK3 does complete prints, they are amazing! But the lack of a stable Octoprint, and the goofiness with the sensors has been a pain.
For fun, I reworked the Einsy base to allow easy access to the RPi Zero. Nothing fancy, but easy to access, and doesn't obstruct the Y-axis stepper.
https://www.thingiverse.com/thing:2734810
Re: Rpi Zero & Octopi working!!!!
Did you install and try latest FW? Supposely they fixed that...?
Thanks for pointing that out, I hadn't spotted RC3_b138 yet. I installed it last night, and as others have stated it no longer crashes at the beginning of the print job. I was finally able to get through one full "Print from SD card" job via OctoPrint, so that's a major improvement. I have not tested uploading a new print job via OctoPrint, will give that a shot next. Have not yet noticed the weird filament sensor behavior or unexpected movements as scott.w describes, but I will keep testing and report back.
Re: Rpi Zero & Octopi working!!!!
Did you install and try latest FW? Supposely they fixed that...?
... Have not yet noticed the weird filament sensor behavior or unexpected movements as scott.w describes, but I will keep testing and report back.
I get those same unexpected head movements where it shoots off to x=0 or y=0. The serial link is still very poor with lots of errors, so I imagine this is because of a bogus move command that is picked up from the bad link.
Using Octoprint just is not reliable yet, particularly bad on an internal Pi but also fails on an external one using USB. The only way to get reliable printing with the Mk3 is through the SD card, it's just not trustworthy using Octoprint. It could also be dangerous as the printer gets stuck and keeps heating and pumping out hot filament until the head is encased. (I’ve had that happen twice now). Until Prusa fix this I would not recommend running the Mk3 unattended using Octoprint. Stick to good old fashioned, reliable Sneaker Net! 😉
Re: Rpi Zero & Octopi working!!!!
Have done several rounds of testing with the bundled PRUSA logo model. Definitely seeing the random movements to X and Y = 0, as well as jerking and stuttering throughout the print. This happens with the OctoPrint 'Upload' and 'Upload to SD' features, as well as OctoPrint initiating a print of a file already resident on the printer's SD card. With the file already loaded on the SD card, it definitely seemed to work better at first. I was able to get 1 successful print that way, but the second time it started exhibiting the same behavior (albeit less severely). And I haven't gotten another clean print that way since.
Still a lot of work to be done here apparently. Really unfortunate, as I was really hoping to start using the time-lapse feature at the very least, but I can't find a way to manually start/stop the recording. It seems you have to initiate the print via OctoPrint for the time-lapse to activate. Anyone know a way around that?
Re: Rpi Zero & Octopi working!!!!
Followed instructions given in the original post and tested with firmware 3.1.1 b138 RC3.
My Pi Zero W is built-in using the recommended frame and no modifications for external access.
I did plug in a USB OTG cable to the Pi, which lets me connect a webcam, rather than fiddling with extending the Pi camera cable.
The serial connection appears to be unreliable. First off, I tried to move the head around a bit in OctoPrint's 'Control' tab.
This worked for a few steps, but then I noticed there was garbled text on the printer's display, so a gcode line must have been messed up.
When trying to print, things started off fine, but while doing infill on the second layer, the head suddenly slammed into the right stop, then continued filling, albeit at an slight X offset. I'm in stealth mode, so no crash detection. I let it continue for a bit, but after a few seconds, it slammed into the left stop, then continued again.
In the 'Terminal' tab of OctoPrint, I saw lots of error messages:
Send: N2524 G1 X101.968 Y110.371 E0.19166*88
Recv: Error:checksum mismatch, Last Line: 2523
Recv: Resend: 2524
Recv: ok
Send: N2524 G1 X101.968 Y110.371 E0.19166*88
Recv: ok
So clearly, serial data gets corrupted. I don't know yet if that's on the Pi's side or the Einsy's side. My bet is on the latter.
The checksum isn't foolproof, so I suspect the erroneous moves were caused by a corrupted gcode line where the
checksum happened to match.
I was previously using OctoPrint on a Pi2, connected through USB. That worked fine.
For now, I have disconnected my camera and am using the OTG port to connect to the printer via USB instead.
Of course, that required disabling the second USART, then resetting the printer using the 'X' button to apply the setting.
This setup works fine.
My plan is to try the second USART again while monitoring the serial line to check if the data on the line matches what
OctoPrint is trying to send.
Re: Rpi Zero & Octopi working!!!!
Without rts/cts pins hooked up and working the best option would be to teach the firmware to use xon/xoff and enable it on the pi as well. That should give a lot better results for people seeing line corruption.
Re: Rpi Zero & Octopi working!!!!
Arduino uses an interrupt driven system for the serial port. So does the Marlin motor firmware. If those interrupts take too long to finish then the UART can get overwritten with the next character before the prior one is read, which I think is what is happening here. Could be related to trying to run both serial ports at the same time. I suspect Prusa might have to disable the USB port while trying to use the internal Pi.
Re: Rpi Zero & Octopi working!!!!
Hi,
Has anyone tried a female and mail headers to mate the rpi zero and Einsy board? Will it still fit case. I'm just not a huge fan of permanently soldering them together.
Re: Rpi Zero & Octopi working!!!!
Hi,
Has anyone tried a female and mail headers to mate the rpi zero and Einsy board? Will it still fit case. I'm just not a huge fan of permanently soldering them together.
It already is that way! 🙂 There is a female header already on the Einsy which the Pi Z plugs into. The only soldering is for the pins on the Pi, not between the two boards.
Re: Rpi Zero & Octopi working!!!!
To be honest, i was a bit surprised that control board on MK3 is still with atmega controller. I was kind of expecting replacing it with arm...
I worked somewhat with UART on AVR controllers and i also had problems with consistency. I hope that Prusa will manage to solve it...
Re: Rpi Zero & Octopi working!!!!
It already is that way! 🙂 There is a female header already on the Einsy which the Pi Z plugs into. The only soldering is for the pins on the Pi, not between the two boards.
Opps sorry did not notice the header on the other side
Re: Rpi Zero & Octopi working!!!!
I tried running with just one serial port and baud rate dropped to 57,600. Much better, still the occasional bad checksum, but just about usable.
I fear, as mentioned above, that the processor is borderline powerful enough to manage all the new features in the main loop as well as a serial interface. This is a problem Marlin had with their main release about a year ago. The main motor interrupt routine was taking so long to run that it left no time for the serial communication. I suspect it will take changes elsewhere in the code, optimizing everything, to solve this.
Re: Rpi Zero & Octopi working!!!!
I had previously been working on mounting a Pi Zero and camera to the left side Y axis pulley. I may go back to that idea, and wire the Pi into to Einsy, rather than directly. Might be cleaner than splicing together camera cables. Would also make the Pi much easier to access.
This was the route I took. It seemed much easier to gracefully maneuver a bundle of six 22awg hook-up wires from the Einsy to the front of the printer (rather than attempt the same with the RPi flat ribbon cable + coupler).
Thanks to scott.w for the detailed writeup. Basic connectivity is working fine for me, however printing via OctoPrint fails almost instantly just as others have described. Hopefully the PRUSA team is working on a firmware fix for this. The MK3 has been advertised as "OctoPrint Ready" since September.
HaHa!
That is exactly the tact I was going to take, it just makes sense dealing with the FFC!
Was going to wait until everything was sorted before I set it up however.
They make a 30AWG silicone wire I was planning to bundle for a super slim EINSY header cable.
Re: Rpi Zero & Octopi working!!!!
Has anyone tried octoprint from a Windows or Mac machine? Does it behave the same?