Strange behavior, printing a bit random...

My Prusa Mendel is acting a bit random and skips a corner on my cube (the 4th), leaving a unstable pattern of strings.

As far a I can tell, this isn't anything mechanical, and not something in the g-code eather, so... what is it?

Well... every layer starts in the lower left corner in the picture (1st corner), going smooth to the lower right corner (2nd), from there to the upper right corner (3rd) and sometimes between the 3rd and 4th corner it skips directly to the 1st corner again for the new layer. This causing ugly prints on other objects and sometimes with too little plastics on the same places.

It always seems to be when printing straight lines on X axis in negative (-) direction.

I'm using RAMPS 1.2 with kliment-Sprinter-0e91bea
Configuration.h
pins.h
Sprinter.h
Sprinter.pde


Other things I've checked out so far:

Baudrate: I've tried lowering the baudrate to 56700 and 19200 with the same results.

Endstop: Never goes near the endstops, they should not trigger and it always goes to the far left when beginning the next layer

Simulate the same problem: not possible, it doesn't print the exact same way twice

g-code: same g-code getting different results when restarting my print

resetting everything: no matter what I do, rebooting the power, uploading the firmware, reconnecting the printer... every time i've got the same problem, but on different layers.

Stepper Controllers: Nope... tried to swap XYZ... no difference...

Skipping steps: Nope... it doesn't skew and always returning to the exact same possition when starting a new layer.

Edited 3 time(s). Last edit at 10/04/2011 06:44PM by adambenganove.

Hi, a was same problem. thas for me was because extruder is little degrees of tilt. ensure is 100% vertical.
Also check the X-axis backlash and X belt
Sorry about my english. hope help

I suspect that your calibration is slightly off. Either your W/T ratio is slightly off so the layers aren't being squashed together and as the print head turns the corner it pulls off the thread (depends which tool you use to generate GCode). Or Z cal is slightly off and it raises the head slightly too high for each layer.

Another thing you could try is to slow down the print speed so the print head moves slower, it may not pull the thread off the layer as it turns the corner.

Good luck.

JB

---

--
Check out my blog: AdventuresIn3-DPrinting

Could you provide us with the gcode file that was used for that test box print ?

Unfortunately i no longer have that g-code that i used when I was printing that cube. But I can provide another cube that gets even worse. Cali_Cube_av_Tobbe-Kopia.gcode

It doesn't seem to be backlash, if it prints anywere close to what it should, then it's spot on.

The extruder 100% verticle, and even if it wouldn't it couldn't be that eather...

It is not the calibration of w/t, because it doesn't "pull the threads off".... the problem is that y-axis starts traveling before it should, so x-axis and y-axis are moving when only x-axis should be the one moving. I hope you guys understand.

When I watch the g-code, it's obvious that it's only x-axis that should be active. It's as if x-axis say's it's done, so the y-axis starts moving even though x-axis have a little bit more to go.

Well that gcode looks fine indeed.

Could it be that your X is sometimes skipping steps for a reason yet to be determined ?

When you repeatedly manually move the X axis forward and backward the same distance - if possible at the same speed you're printing - does it produce that behavior ?

Shouldn't be that as it only happens after the 3rd corner... Could you swap your x and y steppers to eliminate the possibility of a bad motor, as you did for the drivers ?

After that maybe try to find someone who could let you try his electronics - I'm a bit lost about what could lead to that result.

I know you say it is nowhere near your endstops, but is that not the effect you would get if the x endstop was being triggered momentarily during a negative x move? You could try it out by triggering the endstop manually while it was printing.

You've got a good point there... I could try remove the endstops and the problem shouldn't happen at all.

However...

I tried going through the g-code manually to see if I could sort out where exactly the issue lies, and if I could make it get triggered whenever I'd like it to. And here's the fun part... If I made two identical g-codes and on one I removed every E-command (extruder) and boom... BIG difference. (Later this issue have been worse then the picture) So then I tried pretty much the same thing, g-code with E-command but instead I removed the wire to the extruder stepper motor. And that fixed the problem too.... so... I soldered on new connections on both the e-stepper motor wire and the x-axis stepper motor wire. And the issue now is somewhere around 2% of all prints (like the picture above)

So If i do what you say and remove or maybe change the connection on the x-axis endstop, and that should... should solv the problem (hopefully). But once again... good very good point. Thanks a lot!

Nice that you sorted it, but just in case : This could be also caused by an overheating motor. If that would be the case you'd need to reduce its current.