my second raspberry has just passed away

arhi

@spllg said in my second raspberry has just passed away:

so i'm going to wait until one is available before buying s3d

prepare for a loooooong wait (and before you decide to purchase, check their forum and their user experiences first, just to not be surprised later like some)

While I'm sure there are plenty of reasons to carefully consider purchsing S3D, getting prints to a(ny) Duet isn't one - I have used the curl commands with a Duet2 before, and getting it to work with Duet 3 can only be a matter of variable expansion and the correct use of quotes. Worst case scenario we'll have to make a wrapper script, but if it works for curl i don't see why that should be neccecary.

admittedly using ssh is more fiddly unless you want to take the "pull your pants down and hope for the best" approach when it comes to the security aspect of it.

(edit: yes, I'm saying I'll make it work, if it doesn't as posted already)

@dc42 said in my second raspberry has just passed away:

I don't think there is a S3D plugin that supports upload to Duet with attached SBC yet

would i be correct in that its sufficient to check for spaces, qoutes (and newlines?) in a filename to avoid gcode injection if passing it to CodeConsole?

(i.e. S3D calls wrapper with filename as parameter, wrapper sends M32 "filename" to CodeConsole to start print after checking for gcode injection)

gtj0

After copying the file to the SBC, you can call curl to start it printing...

curl -s -d 'M32 "[output_filename].gcode"' http://duetsbc/machine/code

@gtj0 said in my second raspberry has just passed away:

After copying the file to the SBC, you can call curl to start it printing...

curl -s -d 'M32 "[output_filename].gcode"' http://duetsbc/machine/code

That simplifies it a bit, and while I suppose its still possible to mess about with g-code in the file name, its not an additional vector so not my concern.

I'll update my earlier comment to reflect your input

I suppose a HTTP request toPUT /machine/file/{filename}could replace scp as well. Will have a poke later and update the S3D thread.

spllg

going to stay with cura (4.6)

spllg

@spllg

spllg 24 Apr 2020, 08:19

@dc42 thanks for your reply. i cloned from https://github.com/Kriechi/Cura-DuetRRFPlugin/ . this seems to work fine

it seems to have the ability to crash cura - staying with saving gcode to file and downloading this file to rpi.

@spllg said in my second raspberry has just passed away:

it seems to have the ability to crash cura

please consider filing a bug report to the github page for the plugin; it might just help get it fixed by the developers

arhi

@bearer said in my second raspberry has just passed away:

While I'm sure there are plenty of reasons to carefully consider purchsing S3D

like any sw, one should ge informed, I "trusted someone" and didn't, was too late when I figured out who I'm dealing with so I say to everyone just check the forum, if you feel confident about that company and want to be part of that community (where forum is heavily censured, where promisses are never held, where everything is presented as shiny big and huge while... also one important thing I learned from using s3d is to always, always check who I'm dealing with on glassdoor ).
It is still a great slicer none the less!

getting prints to a(ny) Duet isn't one - I have used the curl commands with a Duet2

For RRF3 on Duet2 I use

rrf3.bat "[output_filepath]"

where rrf3.bat is

@echo off

for %%a in (%1) do (
    set filepath=%%~dpa
    set filename=%%~na
    set extension=%%~xa
)   
set if=%filepath%%filename%%extension%

G:\bin\curl-7.50.3-win64-mingw\bin\curl.exe -G --data-urlencode "gcode=M30\"%filename%%extension%\"" http://ender5.local.lan/rr_gcode
G:\bin\curl-7.50.3-win64-mingw\bin\curl.exe  --data-binary "@%if%" "http://ender5.local.lan/rr_upload?name=gcodes/%filename%%extension%"

works like a charm except if file name comes with space then it fails

something similar can for sure be made to work with RRF3+SBC, if you can upload trough web interface, you can send it using CURL ... you put this in the script tab in s3d and forget... it works great

@arhi said in my second raspberry has just passed away:

something similar can for sure be made to work with RRF3+SBC

gtj0 posted the POST endpoint for running gcode and i think the PUT endpoint i added to that should to the trick (i incorrectly assumed it'd be a bigger deal as it wasn't already supported and turned to ssh/scp that would achieve the same thing)

spllg

the 3rd rpi has just passed away.

the printer was switched off for 1 week. after switchin it on it started up fine and i started to print but after ~5minutes it stopped, the http-interface did not respond any more. a reboot of the rpi failed (the green led flashed a few times and was switched off).

i'm tired messig around with rpis (i've got two others and both are up and running for years now).

@spllg said in my second raspberry has just passed away:

the 3rd rpi has just passed away.

the printer was switched off for 1 week. after switchin it on it started up fine and i started to print but after ~5minutes it stopped, the http-interface did not respond any more. a reboot of the rpi failed (the green led flashed a few times and was switched off).

i'm tired messig around with rpis (i've got two others and both are up and running for years now).

I am somewhat puzzled at the point of this post. In an effort to understand the relevance, I need to enquire:

Do you in some way feel that the duet-3 is responsible for the demise of the Rpi?

@spllg said in my second raspberry has just passed away:

the 3rd rpi has just passed away

version of them? and post failure which leds do what, is it the same for all of them? (edit: and how are they powered and conneted to the duet?)

spllg

@bearer Raspberry Pi 4 Modell B; 4 GB, ARM-Cortex-A72 4 x, 1,50 GHz, 4 GB RAM, WLAN-ac, Bluetooth 5, LAN, 4 x USB, 2 x Micro-HDMI

when powering up the red led is switched on. than the green led flashes a few times and nothing more happens.

when i put the sd-card into another rpi it boots fine.

yes, all the rpis passed away the same way - print stopped, rpi does not respond not even to a ping request and does no longer boot up.

the rpi is powered from an original rpi psu, the duet is 5v powered via the ribbon cable (5v->sbc and sbc->5v both closed as instructed by dc42).

spllg

@CaLviNx as it is the 3rd rpi demised in conjunction with my duet3 6hc) while 2 other rpis (not connected to a duet board) live for years now, i feel the duet is somewhat responsible for the deaths.

@spllg said in my second raspberry has just passed away:

the rpi is powered from an original rpi psu, the duet is 5v powered via the ribbon cable (5v->sbc and sbc->5v both closed as instructed by dc42).

so in short the Pi and the Duet are always powered on at the same time? that was mostly what I was after.

understandably frustrating as those things aren't exactly cheap; is there anything else connected to the Pi and have you tried alternative boot medias to the SD card post failure? If you're not interested in spending time and effort on postmortem analysis thats understandable.

spllg

the rpi does not boot from another sd-card (which works fine in another rpi) and the rpi boots fine from the sd-card of the dead rpi.

there's an rpi display and an rpi camera connected to the rpi - unconnecting these devices and the duet ribbon cable does not change anything.

@spllg said in my second raspberry has just passed away:

the rpi does not boot from another sd-card

i was thinking USB or PXE or something not an sd-card?

spllg

@bearer preparing a memorystick using dd. hopefully the rpi is able to boot from usb device.

@spllg said in my second raspberry has just passed away:

@CaLviNx as it is the 3rd rpi demised in conjunction with my duet3 6hc) while 2 other rpis (not connected to a duet board) live for years now, i feel the duet is somewhat responsible for the deaths.

Horse Manure.....

As long as the duet is not sending an over-voltage to the Rpi (and you have said yourself the Duet is not the one sending the Rpi voltage) then I dont see how it is possible for the duet to kill the Rpi.

The logic is flawed i'm afraid.