Print froze… How to troubleshoot?
-
I was printing a set of cookie cutters for my wife (from the sdcard, initiated from the PanelDue), and about 20 minutes in to a 6-7 hour print the extruder carriage just… stopped. No movement on X, Y, or Z, and I didn't hear the extruder motors.
The PanelDue was responsive to touches, but it didn't seem like it was actually talking to the printer anymore. On the PanelDue, I went into "settings" and "save and restart", and the PanelDue seemed to be stuck "connecting..." to the printer. There's an LED on my bed heater, so I could see that the duet was still operating that (flickering LED that's typical with the PID controlled HBP), though I don't know if the nozzle heater was being operated. The web interface wouldn't connect. I waited about 3-4 minutes before I decided to just reset everything.
Later, I realized I had another machine connected to the web interface and checked it's G-Code Console page - and found the following:
[[language]] 10:03:42 PMDisconnected. 10:03:34 PMFile Raptor PLA - R/dog cookie cutter.gcode will print in 6h 37m plus heating time Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. Error: Cannot read file. 9:43:28 PMFile Raptor PLA - R/dog cookie cutter.gcode will print in 6h 37m plus heating timeI'm guessing that the long list of errors is responsible for what appeared to be a lock up. I'm able to confirm (from still camera shots taken every 30 seconds) that the print nozzle did indeed stop at 10:03pm.
At the moment, I'm re-trying the same gcode file from the sdcard, though I initiated it from the web interface this time instead of from the PanelDue. I'll be curious to see if the same thing happens or not. I suppose if it does, there might be an error with file written to the micro SD card (or a specific memory cell on the sdcard.) (Does duet have any write validation switches built in?)
Any suggestions on troubleshooting this?
If possible, I'd like to avoid pulling out the SDcard, as the duet board as that would involve turning the printer on it's side and taking a panel off (and I just got the panel put back on last night.)
Thanks
Gary -
If possible, I'd like to avoid pulling out the SDcard, as the duet board as that would involve turning the printer on it's side and taking a panel off (and I just got the panel put back on last night.)
It looks like you just discovered another corollary to Murphy's Law: securing a panel makes it more likely that you will have to service the item under the panel. How much more likely increases with the amount of work involved in removing the panel.
I discovered another corollary yesterday: the number of times you need to take an assembly apart greatly increases if you use nyloc nuts to secure it instead of plain nuts. It probably applies to nuts secured with thread lock too.
Seriously, with those errors reported, I would change the SD card.
-
Fine.
I just ordered a Transcend 16GB microSD. I chose Transcend as they are the only major memory maker (that I'm aware of) that implements ECC on their microSD cards. 16GB because… well, no particular reason other than it was the smallest I could find (that wasn't overpriced.) I don't foresee needing to store more than 16GB worth of gcode files on the device at any given moment.
The kingston 4GB microSD that came with the Duet Ethernet is formatted FAT32, so I don't expect any issues with the larger card size (and the fact that I can't use FAT16 on a 16GB partition.)
While under there, I'm going to put a short micro USB extension cable on the board, so I won't need to take it apart again if I need access to that port.
Unfortunately, it won't be the last time I have open the bottom, as I still need to find a SSR (not SSR mounted on a logic board) to use for turning on/off the PSU. I also want test how things work with a 40cm 10pin ribbon to my PanelDue (when using type K thermocouples.)
-
@Dc42 I have the same maxim when doing any work on a PC if you close the case side panel before booting it up, the likelihood of your repair being successful drops by 50%, if you do up the thumbscrews you're looking 10% at best.
-
I discovered another corollary yesterday: the number of times you need to take an assembly apart greatly increases if you use nyloc nuts to secure it instead of plain nuts. It probably applies to nuts secured with thread lock too.
Absolutely! I don't know if you are still into flying, but if you take a look at some of the nuts and bolts on an aircraft, you'll find they are further secured with locking wire. The nuts have a hole drilled through across the hexagon. Locking wire is passed through the hole then through an adjacent drilled nut or other convenient point then twisted with a special tool. It's pretty well guaranteed that once all the wire locking is completed, one or more of the nuts will have to be undone
-
…
While under there, I'm going to put a short micro USB extension cable on the board, so I won't need to take it apart again if I need access to that port.
...If you mean a microSD adapter cable, to bring the slot to an external location, then you should know that the documentation advises against that.
-
@bot:
…If you mean a microSD adapter cable, to bring the slot to an external location, then you should know that the documentation advises against that.
No, just the micro USB connector. I don't foresee needing to access the microSD card very often (once I have a working one in there), but I do see wanting to access the micro USB connector (mostly for powering the duet board without having to power up the entire printer.)
