glitched 20 hours into print - random or problem with gcode?
This is from a model I made in openscad, and sliced with prusaslicer.
20 hours into the print everything shifted 4mm left on the X-axis. Is this likely to be a random printer glitch or could it be an error in the gcode? (or a dodgy SD card - although it is a card I've been using for a few months)
RE: glitched 20 hours into print - random or problem with gcode?
Nice looking design.
If it is a good quality SD card, I doubt that it is the card. I have put cheap ones through the washing machine and drying with no issues.
A possible power spike during the print job? A drive pulley coming loose?
You can always pull the GCode file off of the SD card and look at it in Prusa Slicer to see if there is anything there. I learned today to make sure that I look at my gcode closer before printing.
Intermittent problems are very hard to trace.
RE: glitched 20 hours into print - random or problem with gcode?
It looks like the nozzle bumped into some filament standing up due to the many retracts at overhangs. Please take a closer look at the left leg of the arm in front. It is bent slightly upwards. I have seen this before. Especially, if the nozzle then comes from the left, it may even lift the leg a bit which makes things worse.
In case you use normal-mode for the steppers, the controller apparently did not recognize the crash. It might actually have observed several more crashes. If so, it was able to recover those by rereferencing the axis (x?). If you observe such shift, please have a look into the menu where the number of recognized crashes during the current print are tallied.
The cooling fan shroud of the MK3S+ blows from the front and sides. Therefore, overhangs on the back often lack cooling.
With so many near vertical unbraced columns warping is an ever present risk. It appears that one or more crashes has caused a layer shift. Gcode errors and glitches are unlikely to be the cause.
You don't mention your slicer settings, for this sort of print a larger layer height - 0.2mm or even 0.3mm will give you a better chance of completion. Slowing the print and making certain there are no drafts or uneven cooling will also increase the chance of success.
Cheerio,
RE: glitched 20 hours into print - random or problem with gcode?
Ah! that's useful - checked the gcode - it is fine, and recently gave the printer a good service so still unclear....
RE: glitched 20 hours into print - random or problem with gcode?
With so many near vertical unbraced columns warping is an ever present risk.
Once the print gets past the first X, and even more so once past the second X, the columns a pretty solid, and I am using .2mm already.
RE: glitched 20 hours into print - random or problem with gcode?
I'm with the others. If you can exclude a belt issue or loose grub screws, this is most likely caused by the nozzle hitting something warping up.
Formerly known on this forum as @fuchsr -- until all hell broke loose with the forum software...
RE: glitched 20 hours into print - random or problem with gcode?
let me ask you if you have collision detection enabled?
having about 8 years of experience with prusa machines and about 6 years on the MK3 series this is why i disabled all my collision detection settings.
collision detection still does not work properly on my prusa XL and i have disabled there as well.
what happened was the machine detected a collision as the model started to puck up a little as the bottom cooled.
once detection goes off it then rehomes and 50/50% it layer shifts because it screws up on a long print in finding were it actually was when the problem happened.
in the worst case scenario with layer shift or collision turned off you will get a little blob or some other defect which can be sanded or fixed post process.
the nozzle being hot will simply bulldoze through an area of the model that might have lifted and just continue on its merry way.
i know many people have had no issues with the the safety features on these machines but in my experience disabling everything means less problems AS LONG AS YOUR BED PREP IS GOOD. first layer adhesion is most important!!!
if you have parts popping off the bed you have to address that first. everyone has there own procedure for this.
“One does not simply use a picture as signature on Prusa forumsâ€
RE: glitched 20 hours into print - random or problem with gcode?
I've had this problem where prints have a large solid base, but in this case there is absolutely no sign of any warping, everything is still attached to the bed and there are no tell tale lines anywhere to suggest warping.
I'm with the others. If you can exclude a belt issue or loose grub screws, this is most likely caused by the nozzle hitting something warping up.
RE: glitched 20 hours into print - random or problem with gcode?
Not warping from the bed but one of the overhangs on top might have warped upward for the nozzle to hit it.
Formerly known on this forum as @fuchsr -- until all hell broke loose with the forum software...
the columns a pretty solid,
We're not talking about the more usual horizontal warping fighting bed adhesion. Sloping structures, especially thin struts, are known to warp upwards due to having a stick-to surface on one side only. This puts the previous layer in line for a crash...
and I am using .2mm already.
OK, try 0.3mm draft and let us know what happens.
Cheerio,