Bulge when print reaches "solid" layers  

RE: Buldge when print reaches "solid" layers

Thanks Bob!

Interestingly, clicking the link doesn't download the file: rather, it OPENS it. Not as a ZIP, but as a binary... lmao.  And no, I don't know how to wrap a ZIP in a post so it behaves like a zip .

This new forum is a train wreck.

RE: Buldge when print reaches "solid" layers

Thanks, Bobstro, I'll give them a spin on my printer.

RE: Buldge when print reaches "solid" layers

Posted by: tim.m30

Interestingly, clicking the link doesn't download the file: rather, it OPENS it. Not as a ZIP, but as a binary... lmao. 

Well, right-click, download works just fine.

RE: Buldge when print reaches "solid" layers

Posted by: vojtěch.p6

Posted by: tim.m30

Interestingly, clicking the link doesn't download the file: rather, it OPENS it. Not as a ZIP, but as a binary... lmao. 

Well, right-click, download works just fine.

Of course it does if you already know the zip won't download like they do in most any other forum.  What I am describing is similar to clicking an image and instead of seeing the original you get 1's and 0's ... it is a usability thing.

RE: Buldge when print reaches "solid" layers

Well - my first attempt to upload an image that is meaningful to a thread failed with "IMG_0740.JPG Sorry, you are not allowed to upload files." and worse it kicked me out of the post and out of the topic. lmao.

So I am trying again, but using the link method: waiting ... waiting waiting ... nothing.

So now the ATTACH option: umm ... filename is showing but no clue where in the text it will go. So hitting REPLY to see what happens.

RE: Buldge when print reaches "solid" layers

lol - what a let down ... now going back to edit and see if an edit messes things up... well, at least, no major upset there.

No images is sad for a information sharing forum.

RE: Buldge when print reaches "solid" layers

LMFTFY:

(yes, the forum seems pretty broken, but at least you can include your own attachment as a picture if you use the attachment URL while inserting the picture)

RE: Buldge when print reaches "solid" layers

Voj - not sure exactly how you are accomplishing the image insertion, but it is NOT working for myself and a few others... we get a DENIED when attempting to insert an image.

RE: Buldge when print reaches "solid" layers

Yes, that arcane method. Not patricularly convenient, but perhaps good enough until the forum gets fixed.

RE: Buldge when print reaches "solid" layers

Posted by: tim.m30

Voj - not sure exactly how you are accomplishing the image insertion, but it is NOT working for myself and a few others... we get a DENIED when attempting to insert an image.

I had reluctantly started uploading files as attachments, preferring to use an external hosting site for my images (imgur) since I already use it for reddit. In hindsight, I'm glad I did as those images appear to have survived the transition intact. Inserting externally-linked images seems to work better now than uploading images to the Prusa site.

It saddens me to think on how much valuable contributed information seems to have been lost. I'm glad I started keeping my external online project notebook and referring to those pages rather than investing heart and soul here. 

RE: Buldge when print reaches "solid" layers

Anyway, my first print of the test samples. 0.4mm nozzle, 0.2mm "SPEED" profile, Prusa (Filament-PM) black PETG. From left to right 1.35mm to 2.00 mm wall thickness.

I observe three phenomena:

1. Print direction-based thickness variation around the hole.
2. Thickness change when moving from the infill area to the solid wall area.
3. A line bulge corresponding to the solid infill layers.

Effects 1 and 2 seem to be dominating when wall is narrow, but disappear with increasing wall thickness. Effect 3 seems almost constant across wall thicknesses. Biggest unevenness of surface is at 1.35mm wall, where the difference is 0.24mm mainly due to effect 2. Effect 3 at 2.0mm wall size is 0.09mm.

RE: Buldge when print reaches "solid" layers

Posted by: Vojtěch

[...] I observe three phenomena:

1. Print direction-based thickness variation around the hole.
2. Thickness change when moving from the infill area to the solid wall area.
3. A line bulge corresponding to the solid infill layers.

Effects 1 and 2 seem to be dominating when wall is narrow, but disappear with increasing wall thickness. Effect 3 seems almost constant across wall thicknesses. Biggest unevenness of surface is at 1.35mm wall, where the difference is 0.24mm mainly due to effect 2. Effect 3 at 2.0mm wall size is 0.09mm.

I've attached an AMF with the 1.35mm part sliced with my "quality" settings. I found that tweaking the Prusa profiles yielded far more consistent results across all wall thicknesses, and that the bulge at solid infill layers disappeared. I'd be curious to see how the same settings fare on your setup. My take-away from all of this is that external perimeters are the most important part of achieving overall quality. An extra perimeter layer can make a big difference.

RE: Buldge when print reaches "solid" layers

As a data point Bob, I was able to get a bulge to form at full infill transitions even with constraints on perimeter speed.  I'm still in the camp a novelty in the firmware is having issues with certain microstep actions.  Almost like a prior motion command is leaving a remnant that survives into the Y position tables; and why rotating the part 90 will completely remove some of the artifacts, specifically the offset step seen only in the Y axis. 

@ Volj - if you measure the part, you'll find those "steps in thickness" really aren't changes in thickness. They are offsets.  I cut parts to measure true wall thickness cross-sections at those points to confirm.

RE: Buldge when print reaches "solid" layers

Posted by: tim.m30

As a data point Bob, I was able to get a bulge to form at full infill transitions even with constraints on perimeter speed. 

Interesting. What causes it to re-appear at slower speeds?

I'm still in the camp a novelty in the firmware is having issues with certain microstep actions.  Almost like a prior motion command is leaving a remnant that survives into the Y position tables; and why rotating the part 90 will completely remove some of the artifacts, specifically the offset step seen only in the Y axis. 

I'm very curious why other slicers seem to yield better results even with very similar (though hardly identical) settings. I'm hoping to spend some quality time with Cura 4 and ideaMaker in the near future. I would think that will help narrow down whether we're looking at slicer or hardware issues, or (more likely) odd combinations of the two that manifest themselves in different slicers in different ways. With any luck, we'll be able to document some of the idiosyncrasies of using different slicers with the Prusa printers.

Now that I've spent a bit of time away from this issue, I'm going to look at the part orientation you noted, Tim. I'm trying to think of a part geometry that would trigger these sorts of anomalies. I'm thinking of a large (30mm+ to allow full acceleration) square with either a circular protrusions or indentation on each of 4 thin walls (NESW) to try to trigger problems. I'll be away from my printer, so will try to work something up in OpenSCAD early this week.

RE: Buldge when print reaches "solid" layers

A second print, with my tweaks from previous attempts at printing other objects with a smooth surface:

• Print temperature: 245°C (+5°C compared to default)
• Bed temperature: 75°C (-15° compared to default)
• Fan disabled on 1st layer only (vs 3 layers default)
• Fan speed fixed at 30% for the whole print
• No slowdown on small layers
• Max volumetric speed: 2mm³s⁻¹
• External perimeters first
• Seam: Rear

The goal was to force a fixed speed (about 16mm/sec), fixed extrusion rate, fixed cooling and ideally a fixed print direction. This increased the print time to about an hour for a single object. The result is interesting. Effect #1 is still there, as the direction still changes around the hole. Even the increased temperature didn't make the directional texture change (matte vs glossy) go away. But I don't see any traces of effects #2 and #3, at all. Also, elephant foot is mostly gone thanks to lower bed temperature.

So far, for me the biggest change was enabling "External perimeters first", which seems to deal with #2 and #3 almost completely, at the cost of rather ugly visible seams and additional stringing.

RE: Buldge when print reaches "solid" layers

Voj, not entirely sure I agree - I see all three of your numbered effects in that print. The edges and corners are key points to examine. I also see what appears to be almost random seams rather than rear seams - either that - or your filament is wet and sputtering.  What you are describing as directional effects may be simple speed variation due to acceleration time; especially when the texture also changes when the larger area ends, and the Slic3r invokes a different speed in those long strokes along the back, between the hole and the full box layers: you wouldn't be the first person to invert the Slic3r slow down logic..

RE: Buldge when print reaches "solid" layers

@Tim: Yes, I think the extrusion width doesn't change, but rather the trace is shifted outwards.

After having studied how the TMC2130 microstepping and microstepping interpolation works, I don't think it can be related. For identical STEP/DIR inputs, it will provide identical (even if not perfectly linear) motion.

I've also contemplated rounding errors in Marlin. I can't rule those out yet, although Marlin is using floats and so they're not very likely.

Which leads me to think the offsets are generated in the physical world, by the plastic.

I think I should try printing with my two rolls of PA-CF (nylon-carbon), they're almost identical except one is PA12 and the other is PA6/6.6, the first having big thermal expansion (and warping), the second almost none. So we could compare whether thermal expansion is a factor. Given how expensive those filaments are, I'd prefer printing just a single object for each. Any preference on parameters?

@Bob: I'm about to start printing your AMF project now. I'll continue with PETG until the roll runs out, then I'll likely switch to cheap brown PLA.

RE: Buldge when print reaches "solid" layers

The Prusa TMC code uses a modifier to alter the semi-sine microstep correction; if the index to this table gets stepped on or lost and reset an offset will occur.  And it's exactly what we are seeing.  Why only one axis is affected, further hints at something like this.  And, it is speed sensitive. Also - if you look at that photo I uploaded, there is a small box where I was tweaking the TMC, and managed to force a step: but only on one axis: Y. Coincidence?

As for rounding: that doesn't fit the scenario; and I went so far as to select a fixed resolution of 0.2 mm (maybe higher, it was a while ago) without any effect on the print.

Since print speed is the one single factor I have found that eliminates the effect fully, pretty sure thermals aren't involved.

RE: Buldge when print reaches "solid" layers

@Tim:  Interesting idea. However I have Linearity Correction disabled entirely on my printer. Do you see the Prusa TMC code enabling the TMC2130 internal sine table unconditionally? Or do you see the firmware doing the adjustments in software? If so, which line in the code? (I've done a lot of AVR programming, and although Marlin is a bit spaghetti, I can find my way in the code.) By the way, there are two Trinamic microstepping-optimized stepper motors on my desk ready to go into the printer. They should provide a very linear microstepping output even without the sine table corrections. I'll likely install them in the evening (European time).

I'll try to find your post about tweaking the TMC. Hopefully the forum move didn't mangle it.

RE: Buldge when print reaches "solid" layers

@Tim: Ok, I found it. What did you do to induce that step? Also, the TMC2130 microstepping calibration document can be found here: 

https://www.trinamic.com/fileadmin/assets/Support/Appnotes/AN026-Adaptive_Microsteptable.pdf