Bulging corner at layer change - related to start points

[ab3d_at]

Any update on this issue? My calibration print looks the same as the top image of the initial post.

[ururk]

Any update on this issue? My calibration print looks the same as the top image of the initial post.

Not yet. I did a quick test on Friday with Cura and it printed - well not very well, but good for Cura standards (I'm not a fan of Cura). It had similar issues with the corners, but different enough that I don't think I can compare the two.

I decided to try a full-size part, one that normally would take 15 hours on the plus with a .4mm nozzle. It took 5 hours. The part is usable, but had some interesting "defects" at the layer height change which at a smaller scale (.4 width, .2 line height) might not be visible. Plus some banding that always happens when it starts the outer perimeter (but disappears after about 20cm of travel).

I'm going to try and address these other issues first, as they might relate to the square corner issue.

[ururk]

Well, the banding problem is solved, mostly. Still working on it today.

Here's what I changed:

Speeds > Outline Underspeed

Changed it from 50% to 100%

I looked closely at the process, and it is doing a short travel move (no retraction) between the inner and outer layer. Before changing the Outline Underspeed, I was getting a long single-line depression for about 50mm. When I changed it to 100%, the length went to 3mm (barely perceptible).

Check out lines: 2594-2660. As the toolhead moves from 2660 to 2671, something is happening. I'm not too sure if the extrusion is stopping, and in the brief moment pressure is lost, ooze occurs, and the printer has to makeup the pressure loss.

I've attached the factory file.

[ururk]

I think the corner issues are related to layer changes - the bumps are happening when it does a layer change:

[dkightley]

Having looked at the only factory file posted for this thread, I have a couple of comments that might help reduce this issue....

Firstly, I'm used to printing PLA with nozzle sizes of 0.2 and 0.4mm sizes and have little or no issues with the sort of bulging you guys are experiencing.......which may lead me to conclude that the bigger the nozzle, the worse this issue is.

Holding that thought.....looking at the physics of exactly what is happening inside the nozzle, we have a pool of liquid plastic being forced through a hole at a fairly high speed. With a bigger nozzle....and demanding a greater amount of plastic...then the amount of plastic being "shifted" is higher. Stopping the flow of plastic cannot be achieved instantly.....so to get the flow to pause at the right time, then the moving force needs to be withdrawn sooner. Comments on the thread indicate that this "advanced stopping" of the flow has been turned off (or not used). I would prefer this to be tweaked, not turned off!

Now with a larger nozzle, there is a larger mass of plastic being extruded for each layer. With PLA, the sooner the plastic is solidified the better....and with a greater volume of plastic to cool, then more cooling is required. In the factory file, the cooling fan speed is set at 60% from layer 2. I think you actually need 160%...which is not achievable, so 100% will have to do....and for the print speed to be slowed down slightly to give time for the excess heat to be removed...solidifying the plastic before it can get distorted or moved by further printing.

My final thought is that you're seeing ....and I agree quite rightly.....a visible effect of the z axis movement at the end of the layer. Again, looking at your factory file settings,, you have this end of layer point on the outer perimeter....in exactly the place you don't want it. I would use the "outside-in" perimeter order, rather than "inside-out". This will move the left-up-right movement to where it can't be seen....and you then only have to work on tidying up a left-right-left move!

My closing comment.....think about what is actually going on at the exit of the nozzle and try to understand why the plastic is doing what it is doing.....and then work back through the mechanics to find out what to change. Scouring through gcode and comparing setting values, in my opinion, is not the quickest way to diagnose the problem.

[ururk]

Holding that thought.....looking at the physics of exactly what is happening inside the nozzle, we have a pool of liquid plastic being forced through a hole at a fairly high speed. With a bigger nozzle....and demanding a greater amount of plastic...then the amount of plastic being "shifted" is higher. Stopping the flow of plastic cannot be achieved instantly.....so to get the flow to pause at the right time, then the moving force needs to be withdrawn sooner. Comments on the thread indicate that this "advanced stopping" of the flow has been turned off (or not used). I would prefer this to be tweaked, not turned off!

I have tried both coasting and wiping (multi-process to test what effect it had on the defects), neither helped. In fact, coasting caused a gap to appear, and wiping didn't do anything.

Now with a larger nozzle, there is a larger mass of plastic being extruded for each layer. With PLA, the sooner the plastic is solidified the better....and with a greater volume of plastic to cool, then more cooling is required. In the factory file, the cooling fan speed is set at 60% from layer 2. I think you actually need 160%...which is not achievable, so 100% will have to do....and for the print speed to be slowed down slightly to give time for the excess heat to be removed...solidifying the plastic before it can get distorted or moved by further printing.

The reason I turned down the fan - it was causing the temperature of the tip to drop. I have been contemplating designing a different fan shroud so it blows less on the tip, and more on the part. I upgraded my plus (at home) to the 50mm blower fan and printrbot's 50mm shroud design, and had to do the same thing (turn down speed).

My final thought is that you're seeing ....and I agree quite rightly.....a visible effect of the z axis movement at the end of the layer. Again, looking at your factory file settings,, you have this end of layer point on the outer perimeter....in exactly the place you don't want it. I would use the "outside-in" perimeter order, rather than "inside-out". This will move the left-up-right movement to where it can't be seen....and you then only have to work on tidying up a left-right-left move!

I will give that a try again - I had tried switching perimeter order with the test cubes. This becomes a challenge with extreme fillets on the bottom layer of this particular part (the attached factory was just a slice), but I could reduce those.

My closing comment.....think about what is actually going on at the exit of the nozzle and try to understand why the plastic is doing what it is doing.....and then work back through the mechanics to find out what to change. Scouring through gcode and comparing setting values, in my opinion, is not the quickest way to diagnose the problem.

I'm probably going about it backwards, but I'm trying to correlate behavior within the file with a defect. The main reason I was pointing out the gcode lines was because the problem is happening at a specific point - looking at another test print, confirms a bump appears whenever the extruder switches to a fast travel move, which I'm hoping will help point out where in the settings the problem is. Ideally it would give me a better understanding of what the nozzle is physically doing, but I often attack a problem from outside in (if that even makes sense ).

However, I'm probably done for the day so will think about this, maybe read up a bit on nozzle dynamics, and try to think through what it might be doing.

[ururk]

Hi Doug,

Ugh, sometimes I need a boot to the head to reevaluate a situation and look at the bigger picture.

I changed two settings, which have fixed the issues I was seeing:

1) Outer perimeters first. This fixes the z-lift defect. However, when it did its rapid move the perimeter remained broken regardless.

So, after reading through the extruder settings one last time (https://forum.simplify3d.com/viewtopic.php?f=8&t=1980), I:

2) Added a wipe of 5mm

While I believe I tried a wipe before - I think there were other issues with the factory file (the one I attached was after a number of test prints) and I dismissed its usefulness.


To be truthfully honest, I should know more about how the nozzles work (fundamentally) since I do a lot of printing at work, and should be more excited about the dynamics, but my excitement for 3D printing is a bit more, shall we say, focused on the final print than it is about the tedious "fundamentals of PLA flow". And fluid dynamics wasn't my favorite class, I much preferred mechanical-based courses.

[ururk]

Here's a picture of where I ended up tonight:

Wipe distance appears to be longer than necessary, but I'm happy with the direction this is headed in.

[dkightley]

....and I'm glad that my input has contributed to the results you now have..

Its not "one step forward, two steps back"........but "one step back, two steps forward".

[lancej]

I am having a very similar problem with a Volcano extruder and large nozzle on a Prusa i3. (I have attached a picture.) I think the issue I am seeing is exactly what was reported and can be solved the same way. Let me know what you think and whether S3D will solve the problem. Here are the details (similar post in e3d forum):

I am using a Volcano with a 1.0mm nozzle and a 0.7mm layer height. I get a very noticeable "seam" or gap in the outer shell. I realize that some of this is normal, but this one is particularly bad, maybe because of the nozzle size.

The print in question is a 25mm "hollow annular" cylinder printed with PLA with no infill and single width shells. Everything looks great except the seam/gap where the perimeters start/stop. You can see the z-step (horizontal) and a significant gap on every second layer.

I have looked at the g-code and watched the print and I see that there are two differences between good and bad perimeters. One is the direction of the shell print (clockwise or counter) and the other is whether the outer perimeter starts immediately above the outer perimeter finish on the previous z-step, or if it starts after completing the inner cylinder on the same z-step.

How it starts is related to how it ends. If it started on the outer perimeter, it will end on the inner. So, at the end of an outer cylinder perimeter it either moves on the same layer to the inner cylinder or it moves up in z to the outer perimeter on the next layer. I think the latter moves are the "diagonal lines" and the former are the gaps.

I have tried many of the usual things (extrusion multiplier, print speeds, temperature/fan settings) but the issue persists. I even hand-edited the G-code, adding delays overlaps, etc., but nothing has worked yet.

I mainly use Cura (in Repetier host) but I have also tried Slic3r for some of it's extra settings. I do not own Simplify3d, but I'm more than willing to buy it if a solution is likely.

Thanks,
Lance