Any way to improve surface quality on a simple "ring" ?  

RE: Any way to improve surface quality on a simple "ring" ?

Quick update: I reprinted it with the correct filament profile, and it didn't change much (it looks a bit better maybe ?)

So, the question remains, is there a way to improve this to a level closer to vase mode surface quality ?

RE: Any way to improve surface quality on a simple "ring" ?

I could be wrong, but that looks like just too much filament is being extruded, and with over extrusion that filament has to go somewhere so it just pushes the walls out.
A quick test would be to use the prusament that comes with the printer, make sure you select the correct filament profile in Prusaslicer and see if the issue remains.
If that does solve the problem, you'll likely have to work out the correct extrusion multiplier for your filament.

If the problem persists then I'd get onto support. 

Your attachment didn't post - it must be zipped or the software will drop it.

There are several possibilities, lets go with the simplest ones first:

@gb160 might be right in which case you could usefully measure (micrometer) the diameter of your filament - three places about a metre apart and then average - if the value is not exactly 1.75mm enter the actual value in PrusaSlicer and it will calculate the correct extrusion.

Try ticking Print Settings >Layers and perimeters > Advanced > External perimeters first

And, as always when trying for precision: try slowing the print down.

BTW, I see no reason to use 0.15mm layers, 0.2mm should be fine.

Cheerio,

RE: Any way to improve surface quality on a simple "ring" ?

Wow, thanks, a lot of things to try here. I will go over these steps as time permits and report back:

• reprint the part with the slicer defaults settings and a trusty roll of prusament PLA (with black friday deal, I have an entire roll of beautiful lipstick red to use :P)
  • This should address @gb160 suggestion, i.e. that it might be wrong/inappropriate extrusion
• if it doesn't solve the problem already (or as a parallel experiment), i will reprint the part with a variation of these settings
  • Print Settings >Layers and perimeters > Advanced > External perimeters first
  • Slower (@Diem, would be as simple as switching to a STRUCTURAL profile ?)
  • Try 0.2mm (not sure it that could be affecting the outcome but that's gonna speedup the experiments)

Also, I've re-attached the project in a zip. Hopefully it will make it through this time.

A few notes worth adding:

• I printed a 10cm diameter / 20cm height cylinder in vase mode (the lampshade itself) with the same white filamement and it came out close to perfection (i.e. no "waviness whatsoever"). Given that vase mode is essentially a single perimeter, it gives me hope about @Diem suggestion of printing the external perimeter first.
• However, in between printing the "perfect" vase and the "wavy" top/base, I installed the advanced filtration system, which probably affected chamber temperature. I don't think it can be the cause, especially as we're talking about ~10 minutes prints, so heat can't accumulate that much, but just wanted to mention it.

Thanks !

RE: Any way to improve surface quality on a simple "ring" ?

To rule out the advanced filtration causing temp issues, try printing the part with the door wide open/lid off  just for a test with your problematic filament.

See p 44 of your 3D Printing Handbook: section  9.1.3. Speed vs Print Quality

Your part is very thin, only room for two perimeters, both exteriors, leaving no room for internal compensation ... unless you are in vase mode the thinnest wall that will reliably print cleanly without very precise calculations is at lest 5 extrusions thick - at least 2mm and preferably closer to 3mm (with 2 perimeters) which gives the slicer somewhere to lose micro-adjustments.  

You have another issue which will arise as you thicken the walls - look closely, your CAD software has exported the cylinder as a great many short, straight segments - this is fine but a perfect print will consequently have lots of evenly spaced vertical 'corners' - easily mistaken for VFA's.  Your CAD probably has a quality setting or a minimum segment angle setting which could be made finer to smooth the curve.

Cheerio,

RE: Any way to improve surface quality on a simple "ring" ?

Summary of tonight's investigations:

1. Tried with the door wide open (chamber never went above 23C): no visible improvements (so, I guess, it's ruling out the AFS)
2. Tried everything the same except "Print Settings >Layers and perimeters > Advanced > External perimeters first" set to true: hard to tell but it might have improved the quality a bit (the pattern is still visible but looks a bit more "uniform")
3. After reading @Diem reply, I did notice the "facets" looking at the part in the slicer, so I was hopeful that it might be it... I then exported the FreeCAD part in STEP format and replaced the STL in the project, which was immediately visible in the slicer (updated project attached): hard to tell but it might look a bit better, at least the pattern changed a bit
4. With my FW freshly updated to 6.4.0, I though I'd test my belt tension (since I never quite trusted by phone mic based one) and found that my gantry was ever so slightly misaligned. So I did re-align it and re-tensioned the belts using the new wizard: hard to tell as well but I think it did improve things a bit although that's also when I noticed that the pattern isn't the same along the edge (more on this later)

Looking at all the parts, and especially the last one, I did notice a few things:

• The pattern definitively isn't the same everywhere. It is ok-ish around the seam, gets way worse (i.e. much more "wavy", see the first picture bellow), then get much better (i.e. almost perfect, see the second picture bellow), and get back to worse, then ok-ish (it isn't symmetrical, though, i.e. the best part isn't exactly at the opposite side of the seam).
• The first 6 layers corresponding to the base (showing on the right side in the pictures bellow) have a distinctly different pattern (supporting @Diem's comment about the wall being too thin, or possibly being speed related)

I still have to try some of the other suggestions:

• Different/trustable filament
• Different speed/profile
• Thicker wall
• Maybe retry "Print Settings >Layers and perimeters > Advanced > External perimeters first" on top of the other improvements (more facets, re-alignment, ...)

That being said, the irregularity of the pattern, the fact that messing around with belt tension seems to had an effect, could it revive the VFA theory ?

Thanks again for all your advice and patience with novices like me 🙂

RE: Any way to improve surface quality on a simple "ring" ?

Looks like a speed issue ... something wobbling as the extruder moves faster. Look for loose screws.  Did you properly torque the belt drive pulleys when you installed them?  Did you use the FLAT-FIRST method?  If not, the pulleys might be slapping back and forth and distorting print moves. Higher speeds, higher energies to work the pulleys around the shaft. Seems a bit off since the moves will me unidirectional, but resonances may induce slap once the motors get to speed.

RE: Any way to improve surface quality on a simple "ring" ?

@Tim, if by the FLAT-FIRST you're referring to step 5.42 of the Assembly guide, then yes. I remember tightening them that way as hard as I could with the tiny allen wrench they're giving us. I guess I can always take them out and redo it but that would be somewhat of a pain to dissasemble 🙂 Did you have any other screws in mind ?

RE: Any way to improve surface quality on a simple "ring" ?

Flat first refers to having both set screws wiggly loose, the pulley should be able to spin freely.  Then snug the set screw that hits the shaft flat, wiggling the pulley so the set screw flat fully contacts the shaft flat. Then tighten to spec.  Only then do you snug then tighten the second locking set screw.

To remove the pulley, reverse the procedure or you will strip the set screws.

If you snugged both set screws before tightening the one on the flat, all bets are off.  It's essential the tip - or face - of the set screw fully contacts the flat on the shaft. If the second set screw is snug before that flat-to-flat contact happens, there's a good chance the screw on the flat is coming in at an angle and will eventually work loose.

The symptom of the screws loosening is you'll be able to feel some play when you try to move the extruder when the printer and motors are powered on.  Installed properly, there should be no play when you attempt to move the extruder.  And yes, if the pulleys are loose you can have play when the belts are properly tensioned.

I'm only bringing this up because it may explain the odd wiggles in your prints.  Nothing else comes to mind that explains your symptoms as well.

RE: Any way to improve surface quality on a simple "ring" ?

Thanks for the detailed explanation !

I do remember this step in the assembly quite vividly (was only 2 weeks ago) and I am fairly positive that I followed a procedure similar to what you described (I carefully followed the assembly guide which instruct to place the set screw on the flat side first, tighten it firmly and only then tighten the side one). I don’t know if it mattered but I remember that I even cared to position the pulley on the shaft exactly like in the picture (i.e. with the side screw on the right side of the flat part). 

I just tested if there is any play of the extruder with the motor turned on and I could not sense any, the extruded and the rail were “stiff” as a rock (didn’t want to apply to much force either 🙂). When trying to move the extruded or the rail, I could hear the motor hum slightly changing pitch, telling me that the force I was applying was probably directly being transferred to the motor. It is also worth noting that I have probably less than 10 hours worth of print time post assembly, so hopefully, screws didn’t have that much opportunity to loosen.

I don’t know if it rules out that possibility completely but my understanding is that it make it less likely so that partially disassembling the core xy assembly to check these might be more of a last resort kind of thing  (yeah.. being a little bit lazy here 🙂)

I didn’t have time to do other tests tonight but I will also follow up on the few other things that were suggested. 

Thanks,