Printer goes crazy and hangs with code from Cura 14.09

[mpatoulachik 0]

Hi,

I have issues when printing models sliced with Cura 14.09 (and 14.07 a well).

The symptoms:

Although the generated Gcode seems ok, the printer goes crazy in some places and goes all the way to some X or Y endstop before continuing or sometimes feeding out 20 cm of filament, feeding it in again and continuing to print.

Sometimes it skips a layer or two and continues printing.

Sometimes the printer stops in the middle of the print with the heat on. It just stops where it is and still says "printing"

See how it goes here:

http://www.dailymotion.com/video/x2949c6_issue-with-cura-14-09-and-ultimaker-firmware_tech

The video starts after it already went to x- endstop and Y+ then Y- then printed "normally" a few seconds.

Since I know the print is doomed to fail I use all my short leftover filament bits, which explain the color changes...

The setup:

I run an Ultimaker original and Ulticontroller, just upgraded with a heated bed kit (but had the issue before). Slicing from Cura 14.07 and 14.09 gives the issue on some models (not all). The problem existed before I updated to the firmware from 14.09. I believe I had the one that came with 14.04 or something like that. Ironically the ultimaker robot is the model I have most problems with.

Tests:

-From the same computer I sliced the same model (UltiRobot) with the same settings with both Cura 14.09 and 13.10 (was still installed)

-I put both files onn the same SD card (2Gb MicroSD from AData)

-When I print the one sliced in 13.10, no problem whatsoever

-When i print the on from 14.09 the printer goes to endstops, winds out and in filament and finally stops with the nozzle in the part and leaves the heat on (not good at all..)

Thoughts:

When I look at the Gcode in Notepad++, Repetier or Pronteface I can't find any command for this weird behavior (no Y0 or E-200 or anything looking like that). The code looks prety normal although it has a lot of tiny segments and many small travels.

Could it be something that makes the firmware go nuts???

Other thoughts?

 

[Daid 306]

This can be the result of a bad SD card, or communication issues on the long flat cables going to the ulticontroller.

The UM2 has some fixes for this, but those have not been backported to the UM Original yet, because I hadn't seen this problem on that machine so far.

 

[mpatoulachik 0]

Ok, I will look at that.

So far I have printed about 10 ultirobots from both versions of Cura (14.09 and 13.10) and it never flipped out on the file from 13.10 but only succeeded once with the file from 14.09.

But looking at the code I can't see anything that can make the printer freak out on just one of the versions; both files have roughly the same size (64-65 000 lines) and roughly the same paths and speeds.

I will try from USB and from another SD card (the original one I got with bthe printer is broken)...

 

[mpatoulachik 0]

You were right!

The SD card in question is causing problems. I couldn't reproduce the issue with my old (512 Mb) SD or printing from USB.

Quite odd that the bad SD created problems with some files and not others.

Amazing that the corrupted data sent to Marlin was still good enough to be interpreted as coordinates and somehow "printed" (except when the printer got stuck..).

I guess there is a special paragraph about this in Murphy's law like "if somethig can go wrong, it will, but in the way that will make troubleshooting the most difficult and counter-intuitive possible"

I would have expected bad SD issues to show up as files impossible to open or not detected card, not these half-random printing glitches...

Anyhow, thanks for the help, I was miles away from suspecting the SD card itself.