FDM printing on a Prusa just gets interestinger and interestinger ...
 

FDM printing on a Prusa just gets interestinger and interestinger ...  

Page 1 / 2
  RSS
--
 -2
(@-2)
Illustrious Member

Looking carefully at the directional nature of infill artifacts, found my printer doing something not easily explained. 0 degrees is the "what the hell is that?" moment, but then looking down at the 180 degree, a second "why the hell isn't this the same as 0?" moment.

Closer inspect (I have a 6400 dpi view) - and beyond the rather random displacement issue of the 180 - why is there a "modulation" on the X moves that isn't on the Y moves? 

This topic was modified 2 years ago 2 times by --
It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 08/07/2019 8:27 pm
vintagepc
(@vintagepc)
Noble Member

Are you referring to the fat overextrusions in 0 and 180? Those happen because the rectilinear infill sucks at approaching from different sides if the width is not a multiple of the extrusion width, and overextrudes at the places they join. I believe SuperMerill's  p'licer fork has a "filled" rectilinear mod that better handles these narrow spaces in infill. 

You aren't seeing it in the other ones because it looks like it was able to do the infill in a single pass of left to right as a continuous trace and not having to go into a corner and then start/stop again. 

Posted : 08/07/2019 8:35 pm
bobstro liked
--
 -2
(@-2)
Illustrious Member

And further examination proves even more confounding - at least it isn't my imagination I am getting a large L1 variation run to run.

Over extrusion top right pad ...  but just a change in infill direction makes it okay.  LR=0, UR=135, UL=45, LL=90 ...

Exact same part, extra instances, printed a few minutes later, no change in Z-offset ... Lower Middle is 0, which printed just fine 5 minutes earlier. And all pads now showing over extrusion.

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 08/07/2019 11:58 pm
vintagepc
(@vintagepc)
Noble Member

Is your PINDA temp calibrated?

(and is it at or over 35C before the print starts?)

Posted : 08/07/2019 11:59 pm
bobstro
(@bobstro)
Illustrious Member

That 2nd one reminds me of the time I loaded 2 instances of a part and overlaid them. 

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 09/07/2019 12:22 am
--
 -2
(@-2)
Illustrious Member

Sure enough - Temp Calibration, the cure for PINDA temperature sensitivity - no longer works as expected (if at all).  And yes, I ran TC and PID late last week. 

This illustrates same gcode ran with PINDA at 35c and 45c. 

Now to get back to what I was originally trying to understand: the wild spacing variation when infill angle is 180 ...

This post was modified 2 years ago 2 times by --
It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 12:40 am
vintagepc
(@vintagepc)
Noble Member

Maybe it needs to be re-run, some folks have seen corrupted eeprom settings after firmware upgrades, esp. when changing branches, e.g. 3 to 3s

Posted : 09/07/2019 1:57 am
vintagepc
(@vintagepc)
Noble Member

I bet if you rotate the part 90 degrees it'll show up on Y and not X.

Posted : 09/07/2019 2:20 am
--
 -2
(@-2)
Illustrious Member

And this odd modulation on the X-Axis that isn't on the Y-Axis ...

Image 1 with warm pinda, Z-offset @ 700, lin correction Y= 1.060, X=1.050;
Image 2 with cool pinda, Z-offset @ 675, lin correction Y=1.060, X=1.050;
Image 3 with warm pinda, Z-offset @ 688, lin correction Y=1.060, X=off;

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 2:21 am
--
 -2
(@-2)
Illustrious Member
Posted by: vintagepc

I bet if you rotate the part 90 degrees it'll show up on Y and not X.

It's INFILL ...can't rotate it except by setting INFILL rotation, which is what I did. 0, 45,90,135, 180, 270.  Can't get any more basic than that.

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 2:24 am
vintagepc
(@vintagepc)
Noble Member

Yes, the infill doesn't rotate if you rotate the part. - but I think what you're seeing has to do with interrupted infill runs. And rotating the part changes that; if the infill is at 0 but the part is 90 degrees rotated, I think you'll have a one-pass infill run and no artifacts.

Posted : 09/07/2019 2:28 am
--
 -2
(@-2)
Illustrious Member

And the object rotated 90 ... now has a few extra artifacts in the extrusion. Almost like it started extruding from the bottom, then finished a gap from the top, causing the collision high spot.

And It this test it's the micro-modulation apparent in each extrusion path that is interesting. It's visible in X moves, not visible in Y moves.  And they are small, less than the extrusion width.

This post was modified 2 years ago by --
It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 2:57 am
vintagepc
(@vintagepc)
Noble Member

I believe those micro-ripples are the equvalent of VFAs on vertical surfaces from X/Y motor nonlinearity, showing in the overextrusion that's squeezing up vertically. They have the exact same spacing pattern. (there's an umpteen page thread on this I bet you've already seen)

https://github.com/supermerill/Slic3r/releases is the fork I mentioned earlier; If you have a go with that and use 'Rectilinear - filled" as your bottom infill pattern, does the "doubling up of infill lines" issue go away?

This post was modified 2 years ago by vintagepc
Posted : 09/07/2019 3:12 am
bobstro
(@bobstro)
Illustrious Member
Posted by: Tim
It's INFILL ...can't rotate it except by setting INFILL rotation, which is what I did. 0, 45,90,135, 180, 270.  Can't get any more basic than that.

Tim, at one point in the other thread you mentioned first and top surfaces. Downloading your 5mm high part, I've assumed you were troubleshooting problems for both. Looking at these examples, am I correct in understanding that these are all 1 layer prints?

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 09/07/2019 4:44 am
--
 -2
(@-2)
Illustrious Member

Bob, it's a disjointed look into a basic function that is likely the root of artifacts at bottom and top layers. Infill seems to pile up when a slice splits a part and the fill terminates at a line from two directions: a collision, if you will.  Dealing with the layer one symptom is for expediency.  

 

 

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 5:54 am
--
 -2
(@-2)
Illustrious Member
Posted by: vintagepc

I believe those micro-ripples are the equvalent of VFAs on vertical surfaces from X/Y motor nonlinearity, showing in the overextrusion that's squeezing up vertically. They have the exact same spacing pattern. (there's an umpteen page thread on this I bet you've already seen)

https://github.com/supermerill/Slic3r/releases is the fork I mentioned earlier; If you have a go with that and use 'Rectilinear - filled" as your bottom infill pattern, does the "doubling up of infill lines" issue go away?

I was thinking the same, but I have modified linear correction value which should modify the VFA effects.  It doesn't, so my initial conclusion is what I am seeing is not the TMC linearity.   Add to this that X-Axis shows the issue very clearly, Y-Axis seems to not show the issue - but needs more testing.

And there's the issue of X-Axis "dithering" I simply can't explain (Batman prints show it best).  I am trying to get repeatable tests for it, too.

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 5:58 am
--
 -2
(@-2)
Illustrious Member

Well - the micro-artifacts on X are spaced 0.32 mm; or 4 motor steps each.  Not u-steps, but 4 full 1.8 degree motor steps.  Could be the driver, could be the motor.  I need to do a bit more testing to verify Y isn't doing the same, but at a lower magnitude. 

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 6:44 am
bobstro
(@bobstro)
Illustrious Member
Posted by: Tim

Well - the micro-artifacts on X are spaced 0.32 mm; or 4 motor steps each.  Not u-steps, but 4 full 1.8 degree motor steps.  Could be the driver, could be the motor.  I need to do a bit more testing to verify Y isn't doing the same, but at a lower magnitude. 

If it is indeed related to motor steps, would the defect not show at 0.32mm intervals regardless of the extrusion width used? 

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 09/07/2019 6:50 am
--
 -2
(@-2)
Illustrious Member

It does ... extrusion widths of 0.42 and 0.40 make no difference. It is simply a odd ball winding or the driver has a weak FET or circuit around the FET; heck, could even be a bad crimp on the cable.

It is always wise to get more than one opinion... as for trusting Prusa? No way man....
Posted : 09/07/2019 7:20 am
vintagepc
(@vintagepc)
Noble Member
Posted by: Tim
Posted by: vintagepc

I believe those micro-ripples are the equvalent of VFAs on vertical surfaces from X/Y motor nonlinearity, showing in the overextrusion that's squeezing up vertically. They have the exact same spacing pattern. (there's an umpteen page thread on this I bet you've already seen)

https://github.com/supermerill/Slic3r/releases is the fork I mentioned earlier; If you have a go with that and use 'Rectilinear - filled" as your bottom infill pattern, does the "doubling up of infill lines" issue go away?

I was thinking the same, but I have modified linear correction value which should modify the VFA effects.  It doesn't, so my initial conclusion is what I am seeing is not the TMC linearity.   Add to this that X-Axis shows the issue very clearly, Y-Axis seems to not show the issue - but needs more testing.

And there's the issue of X-Axis "dithering" I simply can't explain (Batman prints show it best).  I am trying to get repeatable tests for it, too.

This doesn't alter their position, only their intensity as it modifies the sine wave function driving the stepper slightly. I'm still pretty sure it is VFAs.

Posted : 09/07/2019 11:41 am
Page 1 / 2
Share:

Please Login or Register