The printing time estimate is wrong on both 3.5 and 3.5.1. I started a print Friday that was supposed to take 1d 20h 47 mins. Sliced in both win8 and win10, on Cura 3.5 and 3.5.1, printed using UM PLA white and PVA support, standard settings with 0.06mm layer height. Slicing always gave me the same estimate on both computers, but it took about 5 hours more than what was estimated.
I cannot post the stl as this is a print for a customer and I'm not authorized to share the file.
First time that I personally see such a discrepancy between estimate and real printing time, I know others have had the problem from time to time. I've had prints, even long ones, take a few minutes more than the estimate, but 5 hours is a first for me.
thanks! I just noticed on the file I thought was good that the nozzle did turn off before the motion was done, but it did not matter since the last several moves only took a ~2 seconds.
Cura 2.3.1 is putting that M104 command in the gcode. I did comment that out and things are working better.
I like the idea of the M400 suggestion, this would come in handy in a lot of other contexts too.
FYI - Changing my CURA version is a HUGE undertaking that I fear going through.
No - all your files do this.... 😏
Marlin has a motion buffer of (up to) 16 line segments. And if printing these buffered line segments takes long enough, this can happen. It's just a coincidence that there are a lot short moves at the end of the other half of your files, so that it doesn't hurt that much...🙂
The obvious solution is to remove the "M104 S0" from the end of the files. An additional "M400" right before this line would also help.
Where does the M104 come from? Does Cura 2.3 insert these line or have you added this yourself to the end-script? Cura 3.5 does not add this line (anymore?).