Poor Print Quality on Sloped Surfaces That Face the Rear  

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Just a guess, have you tried a new nozzle?

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

No, I haven't. The stats on the printer show the total print time is 4d 3h. How long should the original nozzle last?

Do the MK4S kit nozzles have a defect?

Have you tried to print the model and see if your print has the same problem?

RE:

Is PETG actually sensitive to draft or is it more so that it'll need more cooling? I suspect the latter, especially if issues show only at certain angles around the nozzle (as the OP realized, improving the print shroud). Question I'd ask myself is, what happens if I print with fans at full blast, doors open plus my Rowenta heater as fan. Material quality will degrade (bonding) but accuracy should improve? Note, I don't print a lot of PETG so I may be off here.

>>  for about 4 hours

I'd dry more aggressively, not that it looks like the root cause. Note, exposing hot filament to undried air is counterproductive (the diffusion speed of water in plastic increases dramatically with T but that works both ways). Dedicated heaters avoid that by design, with a food drier I'd take spools from the running drier into a sealed bag for cooldown. But as said, unlikely to be the main cause, just a thought.

RE:

I gave up on the PETG, and it was printed in an enclosure. The pictures are of PLA; sorry if it wasn't clear. The ambient humidity in the house is less than 15% (minimum reportable value by the thermostat). There was very little filament on the PLA spool and after 6-7 hours of drying, there was no loss of mass.

I would really like for someone else to see if the issue is reproduced or just me.

Are there any improved cooling shrouds to try? If so, why hasn't Prusa been using them? (maybe a rhetorical question)

RE:

Sorry, was just skimming over the post. 

With PLA, "maximum cooling" - door open, side panels off, external air circulation - shouldn't be too far off. 

You could try basic debugging with a Benchie. Read up on all the failure modes that can be seen from it.

I'd also check - warning: potential "short term memory bias" from recent discussion on the German-language forum - that the filament moves freely in the PTFE tubes. The 2nd picture near the hole looks a bit odd, but it's more likely that it warped and the nozzle hit the print object.

RE:

I have printed my model rotated 90 degrees counterclockwise (looking from the top), but still at the same tilt, and the artifact by the hole persisted, and the bad layering appeared on whatever slope was at the rear.

I can try printing outside the enclosure; the sides don't come off of the Lack enclosure.

When I first assembled the printer 1 year ago, the Benchy printed perfectly, but I didn't print it with the bow facing the rear.

I just printed the Benchy in the same Prusament Jet Black filament from the bgcode provided with the printer and it looks perfect, apart from the fact that it is matte while everything else I printed are glossy. I don't have a clue what print settings were used to generate the bgcode.

I just printed Bonkers Benchy from the bgcode for the first time and it is not as good of a print as the Benchy. I have no idea of the difference in settings, apart from almost half the print time, which may be the cause of the reduced quality. I don't know the settings.

Both Benchy models were printed inside the enclosure.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

I printed my model outside the enclosure and the rear right corner printed better, but not properly. The artifact by the hole also printed better, but not properly.

The large area on the right side of the first picture is remnants of the brim.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

First of all if you convert your bgcode back to plain gcode (which can be done in the slicer) you can actually extract a list of all the profile settings as they are stored at the end of the gcode file.  

You can also use File>Import>Import Config in Prusa Slicer and then just select your gcode or bgode file and it will load them in for you.  No need to understand what the settings are called that way.  That allows you to see and compare the settings used in any of the PS gcode files.

Secondly, a matte surface is generally a symptom of the filament being cooler when compared to the same filament printed with a shiny surface.  This can indicate that its being printed too fast in many cases.

All the symptoms you have presented look to be fundamentally caused by insufficient cooling for the areas with problems.  Many many things can contribute to that though.  From fans that need the dust clearing out, poorly functioning fan ducts, to the slice choices for a particular geometry even.

For an example of that last point on a sloping surface your slicer settings may include such things as Vertical shell thickness.  In which case slicer adds in more and more solid infill to cope with potential gaps in the sloped area.  This can lead to an excessive amount of hot plastic in those areas.  Then your fan duct/cooling cant cope with the localised heating and so the filament sags etc and causes artefacts.

Another reason you sometimes wont see it for lower layers near the bed but it starts 'higher up' even with a slope with a constant angle is that when the duct is near the bed the air gets blown back up and around so you can actually get a greater cooling effect near the bed than when its printing higher layers.  At that point only the air focussed on a specific area by the duct is actually cooling the part.  Of course your model geometry can also effect this as flat areas can cause the same effect much like other features can potentially block or divert the cooling flow.  With experience and making use if the different preview options you can learn to anticipate many of these problems.

btw I looked at your original attached project file (many thanks for taking the time to attach it btw).  The option for support everywhere rarely is optimal.  For example that organic tree support for the hole is actually doing nothing.  So I'm not surprised you have an issue in that area.  Basically printing an extreme overhang with no support underneath it.

If I was to try printing that part I would orient it differently and use different settings.  Something like in the attached project file.  I don't have a Mk4 so cant check it exactly but the changes I have made to the project settings (in both print and filament profiles) would be my first start estimate.  If problems persisted then I would then simply just slow the printing down and observe what effect that had.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Thank you very much @neophyl for the detailed constructive help. I will try again with the orientation you suggested.

I have also read that a matte finish is due to fast printing. 

I didn't know about the ability to save bgcode to gcode or that you could import settings from another gcode/bgcode file; those are good tips. The gcode file is a text file; the bgcode file is not. I can only assume that was done to make the file smaller?

I do have some questions about the settings you modified that I hope you can explain to me. I hope you are patient don't think I am being a pest. I just want to understand.

Should I slow the printing time using the dashboard (I remember a speed percent setting on my MK3) or some other Slicer setting?

What are the "different preview options" that you mention? I only know I can look at each layer and tool path in Slicer after slicing. Is there something else I should be looking at?

• Is there any particular reason the object is not in the center of the print bed? Is it just where it ended up after rotating it?
• Ensure vertical shell thickness: You already explained that there may be too much plastic, so you reduced from Enabled to Partial. 
• Infill: Why did you change from Gyroid to Cubic?
• Auto generated supports: You already explained that and painted on supports. Why did you add supports to the edge that is on contact with the print bed? The expanded brim held the edge properly. Without the brim, as the part got larger, it didn't remain square with the print bed and began to flop around.
• Enforce support for the first: I imagine you added it because of the definition "This is useful for getting more adhesion of objects having a very thin or poor footprint on the build plate." My experience with supports is that they leave marks. I may try with just the brim and no additional supports.
• First layer density: The link target in slicer doesn't exist on the linked page. Is this setting to make for stronger adhesion of the support?
• First layer expansion: The link target in slicer doesn't exist on the linked page. Is this to increase the footprint of the circle at the base of the organic support?
• Top interface layers: I read the paragraph at the link target, but don't fully understand. You changed the value from 3 to 4. Can you please explain it in different words?
• Enable dynamic overhang speeds: Please help me to understand what this does and why you disabled it?
• Travel short distance acceleration: The tool tip doesn't do a good job explaining its purpose. Why did you change the default 250 mm/s2 to 0?
• Slice gap closing radius: The tool tip states to keep the value small to increase print resolution. I don't understand the bit about cracks. Can you explain this to me better than the tool tip does?
• Elephant foot compensation: You changed my value of 0.3 mm to 0.1 mm. Why? I have experimented and found that I need at least 0.3 mm or greater to prevent the elephant foot. I found this to be true with my MK3S and MK4S. Did you make the change to help with adhesion to the print bed?
• Cooling slowdown logic: You change the default Consistent Surface to Uniform Cooling. From the tool tip, did you make this change because it slows the printer?
• Perimeter transition distance: You changed the default value of 10 mm to 0 mm, but the field is now immutable. How did it become immutable and why did you make the change? From the tool tip, did you make this change to slow down the printer?
• Enable fan if layer print time is below: You change the default value of 17 seconds to 20 seconds. The part is rather small. Would this change make a difference? Would an even larger value be a better default value?
• Slow down if layer print time is below: You change the default 6 seconds to 2 seconds. I can see that this would slow the entire print down.
• Min print speed: You changed the default value 40 mm/s to 20 mm/s. I can see this will reduce the print speed, potentially doubling the print time.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Should I slow the printing time using the dashboard (I remember a speed percent setting on my MK3) or some other Slicer setting?If just checking that speed changes will make a difference then I'd use the speed setting on the printer lcd.
Once you have determined that is does make a difference then you could create a new print profile with lowered speeds. For example I have a few slower print profiles, but for my printer I mainly ended up reducing perimeters more than the infill speeds. I did a bit of experimentation to see what worked for my setup and what I found acceptable.

What are the "different preview options" that you mention? I only know I can look at each layer and tool path in Slicer after slicing. Is there something else I should be looking at?
OK, after you have sliced and you have the layer and tool paths displayed make sure you have the Legend enabled if you don't already. The legend has a drop down selection field. You can change the preview to show other things such as speeds, volumetric flows, fan speeds etc. There are all sorts of extra information available. Also remember that the Legend allows you to selectively turn on/off all the extrusion types as well as displaying travel moves seam etc.

Is there any particular reason the object is not in the center of the print bed?Is it just where it ended up after rotating it?
I originally copied/paste your object 2 give 2 instances. Naturally I moved the second to one side and experimented with that one. I deleted the original before saving to reduce file size and for clarity. Habit I guess.

Ensure vertical shell thickness: You already explained that there may be too much plastic, so you reduced from Enabled to Partial.
Correct. I pretty much have partial as my default on my profiles now since that option was added.

Infill: Why did you change from Gyroid to Cubic?
Also habit I guess 🙂 Gyroid and Cubic have very similar strength ratios. However Cubic generates less gcode. It also doesn't make my printer throw itself around as much like Gyroid does with its constant changes of direction.  

Auto generated supports: You already explained that and painted on supports. Why did you add supports to the edge that is on contact with the print bed? The expanded brim held the edge properly. Without the brim, as the part got larger, it didn't remain square with the print bed and began to flop around.
Interesting. I've printed quite a bit of stuff with similar orientations and while a brim keeps the small contact area on the plater I have had problems in the past with the part 'flexing' at the small area above the brim. Probably because I print larger parts and the mass goes up alot. So I normally add supports to 'cradle' the bottom of the part and stop it moving. If the brim is enough for this part them by all means remove that extra support by clearing the painting in that area.

Enforce support for the first: I imagine you added it because of the definition "This is useful for getting more adhesion of objects having a very thin or poor footprint on the build plate." My experience with supports is that they leave marks. I may try with just the brim and no additional supports.
Yeah its the same thing as painting it on. To be honest I added the painted area first and then when I was scanning the settings I remembered about that option so added it. If I had remembered before hand I wouldn't have bothered with the support enforcer painting in that area.

First layer density: The link target in slicer doesn't exist on the linked page. Is this setting to make for stronger adhesion of the support?
Yes.

First layer expansion: The link target in slicer doesn't exist on the linked page. Is this to increase the footprint of the circle at the base of the organic support?
Again Yes. I generally alter this on a case by case basis. If you are supporting a larger area then the footprint of the amalgamated support is generally large enough to cause no issues. However if you are supporting small areas then this can result in the base of the slender tree being small itself so making it larger can help.

Top interface layers: I read the paragraph at the link target, but don't fully understand. You changed the value from 3 to 4. Can you please explain it in different words?
Again this is a general thing from my experience with supports. Having a thicker support interface layer can actually make it easier to remove that interface layer. This is more useful for larger flat areas though rather than small points using organic. When you are trying to remove the interface layer from flat areas it has a tendency to break up as you are removing it. This leads to more patchy removal. By making the interface layers thicker they have a greater chance of coming off as a solid area. Sounds counter intuitive I know but Ive had supports just detach with barely any force once you get your setting for a particular filament dialled in.

Enable dynamic overhang speeds: Please help me to understand what this does and why you disabled it?
This changes the print speed based on how much of the layer is overhanging the layer below. So vertical wall and your print speed is what you set. Slope the wall outwards and the print speed slows down. Its something Prusa has enabled by default for some reason. This is great in theory, and for some models with really steep overhangs it can help. However in this case with practically 45 degree walls its not needed. It also has the disadvantage that changes in print speed can be visible on your finished print. Those other visualisations on the Legend, try selecting speed and then turning the option back on and compare the slice if you want to see what its doing.

Travel short distance acceleration: The tool tip doesn't do a good job explaining its purpose. Why did you change the default 250 mm/s2 to 0?
I don't remember touching that.

Slice gap closing radius: The tool tip states to keep the value small to increase print resolution. I don't understand the bit about cracks. Can you explain this to me better than the tool tip does?
I'll preface by saying that I think the default in Prusa's profiles is too large. What that setting does is if there is any geometry on your model that closer to another bit of geometry then the slicer will 'fill in' the gap and ignore it for slicing purposes. You may have seen all those tolerance test models that are available ? Like the toaster or the drawbridge etc. Well the default setting on many of those removes the gap. What you get is people then saying 'oh my printer fails at the tighter tolerance'. When its actually the slicer that has merged things together. Its a stupid default for modern printers that are generally more capable than once from a decade ago. So instead of 0.049 I make it 0.005. That was just a arbitrary value I used that was essentially 10x finer.

Elephant foot compensation: You changed my value of 0.3 mm to 0.1 mm. Why? I have experimented and found that I need at least 0.3 mm or greater to prevent the elephant foot. I found this to be true with my MK3S and MK4S. Did you make the change to help with adhesion to the print bed?
Don't know about the Mk4 but on my Mk3 I only need 0.1mm Elephants foot to combat the spread and still have a solid first layer. 0.3 would indicate that you are really squishing it right in. Then again I'm on a smooth sheet on that printer so ymmv. But yes with a thin edge like we have on the part orientation I thought 0.3 would be causing problems as it would reduce it too much.

Cooling slowdown logic: You change the default Consistent Surface to Uniform Cooling. From the tool tip, did you make this change because it slows the printer?
I didn't touch that all all.

Perimeter transition distance: You changed the default value of 10 mm to 0 mm, but the field is now immutable. How did it become immutable and why did you make the change? From the tool tip, did you make this change to slow down the printer?
I didn't touch that either.

Enable fan if layer print time is below: You change the default value of 17 seconds to 20 seconds. The part is rather small. Would this change make a difference? Would an even larger value be a better default value?
I did alter that, initially I set it to a much lower value, 2s accidentally. When I realised I'd clicked on the wrong setting I couldn't remember what it was so I kicked it back up.

Slow down if layer print time is below: You change the default 6 seconds to 2 seconds. I can see that this would slow the entire print down.
No your assumption here is incorrect. It does not slow the entire print down. Quite the opposite in fact. With a value of 6 seconds if a layer printing time takes less than 6 seconds it slows it down so that it takes 6 seconds. Changing that to 2 seconds means if the layer takes less than 2 seconds to print it slows it down. In effect layers that it would be slowing down are no longer slowed. On some things I've knocked that down to 1 second when I wanted to stop it slowing down. You can adjust based on model size if needed. Then check the previews.

Min print speed: You changed the default value 40 mm/s to 20 mm/s. I can see this will reduce the print speed, potentially doubling the print time.
I didn't touch that either.

Given that several values appear to be changed that I didn't touch then I'm guessing there's some issue around Prusa Slicer temporarily installing profiles and the defaults it is using. I don't have a Mk4 so when it loads in a project file (3mf) that uses a profile that is not installed it temporarily adds it.
I **think** that it loads in the defaults as saved on my pc and applies any ones that you modified from your defaults that are stored in your project. It's possible the default profiles stored on your pc are different from the defaults on my pc so its loading in different values. But who knows, the whole profiles implementation is murky.

In regards to speed settings though, when I'm printing I'm aiming to get all the external perimeters printing at a constant rate. This tends to give you the best surface finish. Infill etc and even normal internal perimeters can of course be different. But on the speed preview my goal is usually to see the outer as the same colour where possible. Its not always possible given geometry requirements but its the aim :)I do tend to make slowest speed as defined in the filament profile match the small perimeters speed in the print profile as in practice those something being printed slow will have a small perimeter. Just what I've found through lots and lots of printing.

Oh and the whole Bgcode thing was created by Prusa to make the file size smaller.  See instead of Prusa fixing the abysmal comms speeds to some of their printers they decided to work around the problem by making the file size smaller, at the expense of compressing it to a binary format.  Great in theory but it makes it unreadable in that state.  Luckily they did provide tools to convert it back but if you don't happen to have Prusa Slicer installed on a machine you are out of luck.  It is a valid engineering decision but not one I'm personally a fan of 🙂 

Hopefully I've not missed anything, if I have just say.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Wow, thank you for the detailed explanations. I really appreciate the effort you put into your reply. I never knew about the additional views in Slicer; I only ever looked at seam position. I looked at the Speed view for the project file you uploaded and except for near the top, it is all the same color. I had to revise a few dimensions in my model and applied the settings you suggested, ignoring the ones you said you didn't touch, and my speed color is not close to a uniform color. Some sloping areas, where I suspect the print will be poor, is the color is blue (~8.3 mm/s) instead of green (~41.8 mm/s). There are few other blue places as well. Would you please have a look and tell me which setting(s) are influencing this speed variation?

I have attached my current project file for reference.

I am going to have to look at the Benchy settings. The seams are indented instead of raised. I would like to find a happy medium once I figure out what settings were used in the Bench bgcode. Also, the MK3S Benchy has the chimney seam in a different location than on the MK4S Benchy.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Sorry I took so long to reply, had to wait until after work.  The main thing effecting it is the Print Settings>Speed>Dynamic overhang speed being enabled.  Personally I don't really make use of that given how it effects the speeds but I'm sure others love it.  If you did turn that off then the only thing slowing it down at that point is the cooling settings.  The 'slow down if layer print time is below' setting in the Filament profile.  As the layers are less than 2s to print it slows it down.

Depending on your filament and cooling it may be possible to print it with that disabled (set to zero).  But that depends on your setup.  By setup I mean the whole of the system.  Everything is in a sort of balance.  You change one thing and it can have knock on effects.  So not just your cooling, but also how the filament reacts to those changes, the environment providing more or less passive cooling etc.  Print with the same printer and the same filaments long enough and you can get a feel for how they behave.  Of course it helps if you do a few experiments to see 🙂  Like you could sink the part to leave just the top slow bit sticking up above the bed and slice.  Print it and see of the peak has enough natural cooling to work.  That way you use less filament and time to check.

The one thing I would be tempted to tweak is Print Settings>Support material>Organic supports>Brach diameter double walls.  Id be tempted to reduce that down.  Mainly as the sloping base of the organic tree looks like there is not much overlap on the extrusions.  On the preview you can clearly see through that.  So I'd be tempted to make it use double walls for that area by adjusting that settings.  Then again if those supports have been stable enough for you in your other prints then leave as is.

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

Thank you very much, again. While I do greatly appreciate your rapid replies, your work must come first. I can't believe I missed the "Dynamic overhang speed" setting. I thought I was careful to reproduce the settings as you described; sorry about that. I didn't want to use your project file because some settings were changed that you didn't make.

After disabling the "Dynamic overhang speed", the only blue slow area is the last 6 mm (triangular bit) at the top. Maybe that's OK since the part is getting thin at that point.

The blue part is about 2 seconds/layer and speeds up to about 1 second/layer according to the the control on the right side of the screen. It makes sense considering the "Slow down if layer print time is below" setting of 2 seconds.

I have cut parts in the Slicer to print just some questionable/challenging areas before. I will be doing that with this project before printing the full object, focusing on the right rear where I have been seeing problems, as well as the blue tip shown in the Speed view. I will post an update.

Would you use the same settings for PETG? That is what I was originally considering using to print this part. My V1 design was overall more challenging to print. Maybe I will try it again using the revised settings you recommended, after I apply some other more recent dimensional changes I made since V1.

Here is an image of V1. The curved part didn't do well at all as I had oriented it, so I squared off the curve as in all later versions. I am not sure it will do well even in the orientation you recommended, but I may try.:

RE: Poor Print Quality on Sloped Surfaces That Face the Rear

I don't have good news. I am at a total loss and fit to be tied as to why the print is so poor, but it seems to be related to cooling. The issue only occurs at the rear of the print on an overhang. I have not had problems with vertical edges. Prusa allegedly validates the default profile settings for each given filament (especially their own), so I don't get it.

I printed 3 different "stubs" in 3 different scenarios just to look at the right rear corner of the print. The brim remnants are visible as a reference; the brim is difficult to remove leaving a huge elephant foot behind. In all 3 stub prints, the edges at the front are perfect.

But first, a phenomenon I have been noticing is curling at the rear and is resulting in the ugly edges. At low layers it is worse. As the object gets taller, the curling is gone. Here is a picture I took after stopping the print with your settings inside the enclosure. I feel the claim of "360 degree cooling" is BS. The duct only surrounds the nozzle a little more than 200 degrees. If they were serious, the duct would surround the nozzle.

This stub is with your recommended settings inside the enclosure. The ugly edges are the same edges that curled. The sides are glossy apart from a lighter dull streak parallel to the brim about 15% up the edge away from the brim; in the picture, it appears darker due to the lighting. The dull streak expands and envelopes the ugly edge. There is an aberration at the other end of the streak at the far edge.

The second stub used the same settings, but with the enclosure removed. Both edges are still ugly, but improved. The streak and far end aberration are still there.

The third stub was using the settings I had already defined and without the enclosure. The dull streak is no longer there. The vertical ugly edge (which is the left rear edge on the print bed) is greatly improved, but the horizontal edge is worse than in the second picture.