What's your average print failure rate? What is your new filament process?

[Dan Royer]

I just tried the same print three times in a row, all failures.  Unexpected X offset mid-print.  Don't know what's causing it.
Then I put new gcode on the SD card and my machine insists the card is empty.  *sigh*.
Feeling more than a little frustrated.  I want reliable, consistent, high quality prints and I don't seem to be able to get it.

Speaking of which, what process do you follow when you get a new roll of filament?  I need to improve my technique for previously stated reasons.

Thank you!

I'm sorry I couldn't make it to Science World on sunday.  I was chairing a meeting and it went from 30 minutes expected to 4+ hr actual.

Dan

[John Biehler]

It's always tough to troubleshoot weird gcode (at least I find it that way)...I don't know how many models I've abandoned because I couldn't print them.

As for your SD card, try formatting it in a camera or some other known good mechanism that will purge any gremlins and recopy the file onto it.

I've had a lot of problems with filament lately due to my slicing a model and reprinting it after changing out filament...I guess I like the idea of the gcode being infinitely repeatable but of course, no two spools of filament are the same so I really need to reslice for each swapout. I was thinking it would be helpful to have a database of filament from various suppliers but every machine has different ways of measuring temps and extruding so I'm not sure how useful my data would be to anyone else.

Science World was fun...I had to leave early but lots of good stuff was going on.

[LoialOtter]

If you're using Repetier Host, you can go to the g-code viewer and look at it layer by layer or in layer sections. Use that to double-check that the slicer isn't giving you weird results.

If it is, I suggest using netfabb lite to fix it as that does a pretty good job and is free.

If you aren't using repetier hose, try it out with viewing the code if nothing else, it's a fantastic feature. All the gcode I've thrown at it it's been able to read without trouble including the weird hand-written stuff I have to use on some things (such as my laser cutter). It's also a handy way of fixing up g-code when you need to put an M84 after the first z-move (turn steppers off so you can hand-tweak the z-offset during print) or you want to remove the M104 and M140 at the end to keep the bed and hot-end on after print.

[Dan Steele]

Hey Dan,

I keep a chart of what temperature all my filament prints best at and change the temperature and turn the build platform on/off in the gcode before each print depending on filament type.  I print directly from my computer so this part is pretty quick.  I also keep a spreadsheet of what slicer settings I use for each object that I make, if you're printing things over and over again it's worth it to be pedantic about it.  Once I have a clean model that I've used before successfully my failure rate is quite low and only results from a jammed filament spool, or a miss-aligned build platform.

As for broken models I run into them all the time, there are many broken models on thingiverse, it's also quite easy to make one...  I use netfab sometimes but I find that as often as not it makes things worse.  I often poly reduce in Blender to something reasonable (under 20K polys) then import the stl into sketchup delete all the co-planer edges with a script and then fix the model by hand...I've gotten pretty fast at this and the results are a model that has fewer polys and that you can rely on.

Dan.

[Dan Royer]

When I say failures, I mean machine mistakes.  The gcode looks fine, repetier and netfabb are both happy with it... and then at some random point in the print I'll get a fault line and everything above it will be offset on X or Y and BOO HISS RESTART PRINT.

[Andrew Lepper]

Failures are 1 in 20 for me. Formerly 1 in 5 due to preventable things like accidental USB knocks, filament spool knots, and incorrect z height settings. Still not perfect.

[Dan Steele]

If I use the whole bed  maybe one in 5ish due to my bed being slightly warped,  If I use the centerish 1/20-1/30 somewhere in there..I was getting a much higher rate of failure with ABS before I enclosed my machine.  PLA is much more forgiving..

How fast are you printing?

Do you have accelerated firmware in place?

ABS/PLA?

What's your nozzle diameter?

What layer height are you using?

Dan.