Newb, Printed some tests - Can you help me interpret them?  

Re: Newb, Printed some tests - Can you help me interpret them?

shit thats a lot. your prints look fine, my dude.

layer "banding" : welcome to the great saga of mk3 extrusion inconsistency. stay tuned.
surface text : silver makes it look worse but slow down your infill layers. assuming you're printing in normal mode at full speed
zits: normal restart/retract artifacts. play with coast settings and accelerations settings. not bad if you're at full speed.

edit: on second glance your biggest recurring problem is ringing and underextrusion on the top infill layers.
edit2: also don't ever use 100% infill, it just causes problem. actually there's very few scenarios you need more than like 30 or 40

Re: Newb, Printed some tests - Can you help me interpret them?

hey curtis, thanks for the response.

Yeah, I went all gung-ho with the post. But I though if I were to receive help it would better be a well organized and full of useful images post 🙂

Layer banding - I was about to post a comment (the forum doesn't let me edit my first post) about the layer banding when I saw your response.
Just yesterday I read about the issue #602 on github. Mine has been making it right from the first print (benchy was my second print ever) so I thought it was normal because this is my first 3D printer and I know nothing :D. Now I see it's not.

surface text - I was printing at full speed, yeah (170mms/s for infill acording to slic3r), thought the Mk3 was capable of printing well at 200mm/s. So slowing down the infill, got it. There are lots of settings I don't know when I should use if at all "only retract when crossing perimeters", "infill before perimeters", etc

Also, I printed with 100% infill because the author on thingiverse said to do so, thought it was strange but.... I guess it's because if not the surface where the text is would appear half filled? That's one of my doubts. The pinter considers top faces every top face of every surface regardless of height? or just the topmost faces of the object? I mean, slic3r does 5 top solid faces on every horizontal surface that doesn't have another layer on top?

zits - can you point me in the right direction? can't find those coast settings. Can only see this somewhat related settings, and the retraction related settings are on the printer specific settings. If this are normal and to be expected I think I was spoiled seeing some mk3 review, on Tom Sanladerer one he shows some perfectly regular and smooth prints 😀

So what's the best way to correct those ringing and underextrusion on top infill layer problems?, I've read about calibrating the extruder and e-steps and whatnot, but a lot of people says it's not necessary on the mk3 and that Prusa has already done that.

Maybe I should update to 3.2.0 and configure the linearity correction to see if that does anything. Still wary though because of the problems.

The more I read and learn, the less I know.... 🙄

[edit] Also, is there a way to correct that drooping layers (points of the letters on the XYZ cube, holes on the front of the benchy...) that's a case of not enough cooling because of small object right? should I speed down a lot? or print more than one object to give them time to cool down? any other way? because I see those a lot (the underside and ears of Marvin...)

Re: Newb, Printed some tests - Can you help me interpret them?

3.2.1 firmware is a must on MK3. With linear advance you will get a lot better looking prints.

Overall you have a good start. My advice is to forget about all the test objects. Go ahead and use your printer for whatever reason you bought it. All the tweaking and learning will come with the real usage automatically.

Re: Newb, Printed some tests - Can you help me interpret them?

3.2.1 firmware is a must on MK3. With linear advance you will get a lot better looking prints.

Overall you have a good start. My advice is to forget about all the test objects. Go ahead and use your printer for whatever reason you bought it. All the tweaking and learning will come with the real usage automatically.

Just out of curiosity should I enable linear advance? I updated the firmware as soon as it came out but I noticed this morning that linear advance wasn't on in the setting. is it worth doing or is it something for just specific situations?

Re: Newb, Printed some tests - Can you help me interpret them?

3.2* firmware is handling linear advance command again. It's nothing can be activated or deactivated in the settings of the printer. It's part of the slicer config.
https://help.prusa3d.com/l/en/article/t5w9VsdVai-linear-advanced

You just need to make sure "M900 K30" command is in your start script. That activates linear advance for the printed file. Maybe it's already there in your slicer. Search for a M900 command.

Re: Newb, Printed some tests - Can you help me interpret them?

I'm a idiot apparently I was thinking of E-Correct being off

Re: Newb, Printed some tests - Can you help me interpret them?

thought the Mk3 was capable of printing well at 200mm/s.

capable yes. but physics gets in the way at some point and if you're interested in maximum quality...gotta slow it down.
Isaac Newon >>> Jo Prusa

zits - can you point me in the right direction? can't find those coast settings.

Apparently slic3r doesn't support coast...which is dumb. I normally use s3d but trying to learn slic3r and I'm not a fan of its profile organization system. Printer settings tab will have your wipe/retract/hop settings though. The defaults built into the prusa edition of 1.40.0 alpha 1 work pretty well for me at much lower speed than you're using.

for the other artifacts you're seeing, start with this guide from s3d.
https://www.simplify3d.com/support/print-quality-troubleshooting/

Tom Sanladerer one he shows some perfectly regular and smooth prints 😀
Well Tom is a pro. The only really "fair" comparisons are the presliced demo files (benchy) and yours looks pretty good.

Maybe I should update to 3.2.0 and configure the linearity correction to see if that does anything. Still wary though because of the problems.

3.2.1 is current as of today actually. I'd redo the xyz cal since it was updated apparently.

I just rebuilt my x axis so maybe I'll print a calicat and see what it looks like....

Re: Newb, Printed some tests - Can you help me interpret them?

3.2.1 firmware is a must on MK3. With linear advance you will get a lot better looking prints.

Overall you have a good start. My advice is to forget about all the test objects. Go ahead and use your printer for whatever reason you bought it. All the tweaking and learning will come with the real usage automatically.

🙂 yeah, just thought that specific tests would be better to really know where my printer fails. I have to find something that doesn't took half a life to print until I can see if the changes I make are worth it.

In the meantime, having a 6 y/o daughter i've been having to print tons of calicats 😀

I'm still wary to launch multihour prints because of the "problems". The more complex things I've printed are the gear bearings (success at the second attempt) and the iris box (half success, printed kinda ok but paused mid print once I saw this and I guess some layers are fused in the middle. Looks like it could work, it kinda moves but can`t get it to spin. Seems like over extrusion at the end of those layers or bad cooling... I don't know.

I'll update to 3.2.1 right now, recalibrate xyz (factory reset advised?) and print that infamous ecor tower to calibrate the linear advance.

capable yes. but physics gets in the way at some point and if you're interested in maximum quality...gotta slow it down.
Isaac Newon >>> Jo Prusa

😀 yeah, you're right. I have the printer on a table and although it's a sturdy table I can feel the vibrations on the wood when the printer is going full pelt on big infills. Thinking of printing some kind of dampeners.

What's a good infill speed? (since perimeter speed is already at 45mm/s by default in all the profiles). I mean, it's at 170mm/s by default from the theoretical maximum of 200. But on other printers I see people recomending everything from 50mm/s for everything to 150mm/s for the infills

Apparently slic3r doesn't support coast...which is dumb. I normally use s3d but trying to learn slic3r and I'm not a fan of its profile organization system. Printer settings tab will have your wipe/retract/hop settings though. The defaults built into the prusa edition of 1.40.0 alpha 1 work pretty well for me at much lower speed than you're using.

for the other artifacts you're seeing, start with this guide from s3d.
https://www.simplify3d.com/support/print-quality-troubleshooting/

After your post I read about it, that would be a good setting to have (distance from the end where the extruder stops extruding right?).

Although I'm new to this I feel like Slic3r while being kind of easy to understand , has a horrible UI in terms of organization with settings all over the place and the profile organization is kind of confusing (although it can be that I'm new to this of course). Retraction, wipe, etc is on the printer settings tab so if you want to change it for a print (and store the changes) you have to create a new printer profile, although for what I've read, it's a setting that is somewhat common to change for different filaments, so it could be easier if you could change it there, or at least having a modifier there on the filament profiles?

🙂 Thanks, I have that guide along with this one on my start page now. Sometimes I find it difficult to really differentiate the problem (overextrusion from overheating for example in the images I linked above in this same post) or not being able to recognize or know I have that problem at all (ie rosparovac mentioned my printer may be underextruding, and I don't see it does) so I wouldn't know if I need to calibrate the extruder or change the flow or whatever. I know that'll come with time and prints though. Just want to stablish a baseline of what a good print looks like now that the printer is new and it is at it's peak performance in terms of wear and tear, if that makes sense.

Re: Newb, Printed some tests - Can you help me interpret them?

What's a good infill speed? (since perimeter speed is already at 45mm/s by default in all the profiles). I mean, it's at 170mm/s by default from the theoretical maximum of 200. But on other printers I see people recomending everything from 50mm/s for everything to 150mm/s for the infills

Just depends on how your prints are coming out and what your tolerance for noise and vibration is. I won't go above 120 or so and 50 for the top layer. With gaps you have in the top layers and the details around the lettering, you want to look at the infill overlap and top layer speeds especially.

Although I'm new to this I feel like Slic3r while being kind of easy to understand , has a horrible UI in terms of organization with settings all over the place and the profile organization is kind of confusing (although it can be that I'm new to this of course). Retraction, wipe, etc is on the printer settings tab so if you want to change it for a print (and store the changes) you have to create a new printer profile, although for what I've read, it's a setting that is somewhat common to change for different filaments, so it could be easier if you could change it there, or at least having a modifier there on the filament profiles?

The idea is that you may only want to change filament types for example, so you'd only need to change filament settings. But when would you ever change filament types and not also change retractions, speeds, etc etc.

Slic3r definitely gives off that "free shareware from the 90s downloaded off a geocities blog" vibe.

Re: Newb, Printed some tests - Can you help me interpret them?

Just depends on how your prints are coming out and what your tolerance for noise and vibration is. I won't go above 120 or so and 50 for the top layer. With gaps you have in the top layers and the details around the lettering, you want to look at the infill overlap and top layer speeds especially.

Noise and vibrations are already not that bad but yeah, I know what you mean. By default (and haven't touched those) infill overlap is 25% and Top solid speed is 50 mm/s.

I updated to 3.2.1 and did the linear calibration printing the ECOR tower. Pretty good results (of course is presliced by prusa so...) with "just" two small separations between layers, and, of course those damn gigantic seams that I guess won't ever go away . On the other hand I measured the walls and they are exactly 0.45 so I guess I'm not under extruding.

Just to test the new firmware after the linear advance calibration I reprinted the iris box (it works this time) with the only change of lowering both infill speeds (Infill and Solid infill) to 100mm/s. It works well although on the exterior looks like crap, full of zits and seams, and the closing leaves look horrible.

Finally, while reading the extrusion inconsistency discussion on github I printed some XYZ cubes. The 0.1 one is the default Slic3r PE detail settings and looks absolutely terrible. The middle one is printed with a gcode (downloaded from the discussion) that someone sliced with simplify3d and the one on the right I sliced with Slic3r PE at 0.2 with both infill speeds at 90mm/s and it doesn't look bad apart from the seams and the extrusion inconsistency problems (present in all of them).

Clearly the best one is the simplify3d one. Probably because of my inexperience in setting slic3r, but I'd love to do more tests. Unfortunately I don't own S3D (nor am I going to spend 150 on it atm). Since you use both, could you send me a gcode sliced with both with the settings you have good results at 0.2? A calicat or cube or something. I know there are lots of factors so good settings won't be the same for every person but would put me in the right track wether they work better or worse for me I can look what's changed.

The idea is that you may only want to change filament types for example, so you'd only need to change filament settings. But when would you ever change filament types and not also change retractions, speeds, etc etc.

exactly, not every filament type has the same retraction settings. even i know that. They could keep the main settings on the printer tab as a base for if you have more than one printer and then let them be modified by other settings on the filaments tab

Slic3r definitely gives off that "free shareware from the 90s downloaded off a geocities blog" vibe.

Oh boy, that win98 feeling 😀

If only they applied what they have done with Prusa Control to the full Slic3r interface....

Re: Newb, Printed some tests - Can you help me interpret them?

damn gigantic seams that I guess won't ever go away

thats the default 'wipe' behavior that slic3r uses instead of a coast, plus the fact that its a single walled model. much less noticeable normally.

the alpha version of slic3r finally has a real full featured gcode preview that will be very useful to you to see where those retractions are actually happening.

I'll see what I can do with s3d, currently reassembling my y axis. Need to design a better belt holder because this one blows.

Re: Newb, Printed some tests - Can you help me interpret them?

thats the default 'wipe' behavior that slic3r uses instead of a coast, plus the fact that its a single walled model. much less noticeable normally.

well, I hope they think about implementing that coast in slic3r then since, in that regard, results are much better. With slic3r (and I'm sure it's my own fault, I have seams and seam-holes all over the place

the alpha version of slic3r finally has a real full featured gcode preview that will be very useful to you to see where those retractions are actually happening.

you mean as in graphically? I currently only open gcodes on Notepad++ to see the settings used (because I don't "read" gcode, but it would be nice to see a reconstruction of the extruder movements (I see S3d does that on the preview) unless you're referring to something completely different. I'm weary to use alpha versions at this stage to not introduce more factor into my tests

I'll see what I can do with s3d, currently reassembling my y axis. Need to design a better belt holder because this one blows.

Thanks, I feel like having the same object sliced on the two slicers with good settings (meaning settings that work well for a more experienced user) will let me compare in the same conditions and extrapolate.

I also think there are a few things that I'd love to "upgrade" (toothed idlers, vibration of the z screws, offload the filament to the side instead of on the frame, vibrations on the screen case, dampeners on the legs, better ways to acces and fix the belts tension...) but I want to first learn and tune the stock machine (with Prusa's own filament) to rule out any changes I' dmake.

At least I know my dimensional accuracy is relatively good since I had two "complicated"prints working although the looks are not that good (not taking the extrusion inconsistency into account) 😀

Heeding @nikolai.r at least means my kid will have a lot of little toys to play with.

Re: Newb, Printed some tests - Can you help me interpret them?

Thanks, I feel like having the same object sliced on the two slicers with good settings (meaning settings that work well for a more experienced user) will let me compare in the same conditions and extrapolate.

try this one. change the file extension to *.gcode

Re: Newb, Printed some tests - Can you help me interpret them?

try this one. change the file extension to *.gcode

Boy, that resulted in a mostly flawless XYZ cube, some minimal ringing , some not totally bonded strings in the first layer and some weird things on the top layer (some overextrusion and nozzle "crossing" that - I observed those same exact marks on the other xyz cube I downloaded from the inconsistent extrussion discussion, also sliced with simplify3d). The little stringing I'm guessing is because the nozzle was at 215º the whole time. Even the inconsisten extrusion problem seem greatly diminished or at least not that aparent (though it's still there)

You can see the results here

At first I thought you sliced with Slic3r 1.40 while digging a through the gcode and didn't understand some things in there, now I see you used S3D 😀

Some interesting things I can see/understand, though maybe they don't translate the same to slic3r:

That's as far as I can see. Do you get the same final results in your printer if you slice it with slic3r?

For comparison you can take a look at a Cali mouse I printed yesterdaywith the standard prusa fast profile (0.2 mm). The only change I made was slowing infill and solid infill speeds to 90mm/s instead of the default 170mm/s

Re: Newb, Printed some tests - Can you help me interpret them?

Boy, that resulted in a mostly flawless XYZ cube,

wow yours actually looks a lot better than mine, but I don't have any standard PLA to test with at the moment. I took my 3dfuel APLA+ profile and just lowered the temps for the gcode I gave you.

Even the inconsisten extrusion problem seem greatly diminished or at least not that aparent (though it's still there)
this is also a lot better on yours than mine and many other I've seen.

Some interesting things I can see/understand, though maybe they don't translate the same to slic3r:

That looks like an effect of the orientation you printed in. The extrusion width can only be varied so much by the slicer so its possible to end up with small gaps.

Check out the infill pattern on the tail when you change its orientation 90 degrees. Not sure if slic3r can do this, but I can change infill angles if it becomes a real issue. Changing infill extrusion width is another solution. But either way its not the printers fault.

Re: Newb, Printed some tests - Can you help me interpret them?

[*]Am I reading this right? you are using a 105% first layer height , a 102% first layer width and a 105% infill widht? (I don't think you can change that on Slic3r you actually can, but I don't think these are needed in your case since your bottom layer looks slightly squished. this is a carryover form several firmwares ago when I was having first layer issues

I thought I would point out a common misconception about Simplify3D's "first layer height" setting. Whatever you have this set to, the height is adjusted by that value, but it extrudes the same amount of filament as it would for a normal height layer. Hence, if you have this set to 75%, you actually get more squish as it is extruding the same amount of filament (for a normal layer height) in a layer that is just 75% height of a full layer. If you set it to higher than 100%, then you won't get quite as much squish.

Re: Newb, Printed some tests - Can you help me interpret them?

[*]Am I reading this right? you are using a 105% first layer height , a 102% first layer width and a 105% infill widht? (I don't think you can change that on Slic3r you actually can, but I don't think these are needed in your case since your bottom layer looks slightly squished. this is a carryover form several firmwares ago when I was having first layer issues

I thought I would point out a common misconception about Simplify3D's "first layer height" setting. Whatever you have this set to, the height is adjusted by that value, but it extrudes the same amount of filament as it would for a normal height layer. Hence, if you have this set to 75%, you actually get more squish as it is extruding the same amount of filament (for a normal layer height) in a layer that is just 75% height of a full layer. If you set it to higher than 100%, then you won't get quite as much squish.

you're correct, I've just had better results with adhesion by using a taller first layer layer. Doesn't seem to be an issue anymore since mesh leveling was improved.

Re: Newb, Printed some tests - Can you help me interpret them?

wow yours actually looks a lot better than mine, but I don't have any standard PLA to test with at the moment. I took my 3dfuel APLA+ profile and just lowered the temps for the gcode I gave you.
this is also a lot better on yours than mine and many other I've seen.

I think this comes to show that the inconsistentcy is mostly software/firmware related

[*]you do a 4 perimeters skirt, any particular reason or just personal preference? If I'm not using a brim, I like to give it plenty of time to dislodge any strings picked from oozing during preheat and mesh leveling. This is also why I turn the fans ON during preheat
That's a good idea. I also have to learn how to do proper custom gcode to add a good preheat sequence and raise the z level on print end, it stays too low so It's difficult to remove the bed.

[*]on the gcode your extruder retraction distance is 1.4, is that mm? because the default on slic3r retraction lenght is 0.8 mm never found .8 to be sufficient for me

I always have a small bit of stringing, maybe I should raise the retraction distance a bit, although it also happened with your gcode, maybe it was due to it being hotter than the standard pla temps.

[*]you have 3 top and bottom solid layers slic3r has 4 and 5 5 is just excessive for pla unless you're running really low infill or you need a super strong part

Noted :), it's always good to save a little bit of plastic. My top layers are not good though

[*]Am I reading this right? you are using a 105% first layer height , a 102% first layer width and a 105% infill widht? (I don't think you can change that on Slic3r you actually can, but I don't think these are needed in your case since your bottom layer looks slightly squished. this is a carryover form several firmwares ago when I was having first layer issues

I'm still fiddling with my live z adjustments I had it pinpointed on 3.13 but since flashing 3.2.1 and redoing the calibrations I'm still not happy with my new live Z

[*]haven't found the speed, at least in a way I can understand it[/list]
These are the speeds I'm using
I'm also using very low acceleration and jerk values which I commented in the start script.

So translated to Slic3r speeds that's more or less:

Unless I'm reading that wrong of course, also slic3r seem to have more speeds to control (i.e distinguises between solid infill and top solid infill).
Also jerk and acceleration , I saw them on your start script but I'm not really sure what they're for or where to change them on slic3r (or what values they're set at for that matter.
For comparison, this are the speeds at which I sliced that cali mouse

That looks like an effect of the orientation you printed in. The extrusion width can only be varied so much by the slicer so its possible to end up with small gaps.

Check out the infill pattern on the tail when you change its orientation 90 degrees. Not sure if slic3r can do this, but I can change infill angles if it becomes a real issue. Changing infill extrusion width is another solution. But either way its not the printers fault.

You know what? it never occurred to me that that could be the problem :). You're totally right, and yes, Slic3r allows you to change that angle.

Watching videos I really like the way S3D shows the preview (different colors for inner and outer infills, etc) and the general UI. It's leaps and bounds better than Slic3r, a lot clearer. The preview on slic3r is confusing for a noob like me. Unfortunately I don't think S3D is worth the 150$ at this point (compared to Slic3r I mean), maybe a year ago but not now that I'm starting.
Although it seems to do some things better than Slic3r, your cube doesn't have those small zits and the seams are almost invisible (just a small bulge on one corner)

It would be cool to have it to learn and I'd love to play with it but 150$ is A LOT of filament (of which I have to get some by the way, I'm eating through the included silver pla)

I think we can say at this point that the problems I was seeing when started this thread are mostly due to incorrect speeds for my use case (wether it is because of the location of my printer or whatever, it's clear the default profiles are not totally cutting it for me). I also have some other problems (like the bit of stringing, squished blobs in first and top layers, still some ghosting, and not perfect top layers) but I think they're kind of minor compared to the jump in quality I saw with your gcode where seems like the main difference are speeds.

I don't know that I'll be able to fully "translate" your gcode into a Slic3r profile but I'll certainly try

I thought I would point out a common misconception about Simplify3D's "first layer height" setting. Whatever you have this set to, the height is adjusted by that value, but it extrudes the same amount of filament as it would for a normal height layer. Hence, if you have this set to 75%, you actually get more squish as it is extruding the same amount of filament (for a normal layer height) in a layer that is just 75% height of a full layer. If you set it to higher than 100%, then you won't get quite as much squish.

hey, thanks for the clarification. So if I get it right if you set it higher than 100% you're not only getting a taller height but also less width? I mean, if you use the same amount of filament for a higher height you must get less width, I guess that's why you also have to use a higher than 100% width.
That over compensation may explain what I see as the only flaw of the gcode curtis sent me, the top layer since it looks a bit irregular and shows signs on nozzle dragging.

Not that it's of use for me on Slic3r but it's always interesting learning how things work.

Re: Newb, Printed some tests - Can you help me interpret them?

just taller. 150% would be .3 if you started with .2.

way quicker than the standard 45mm/s on Slic3r PE

Its % of the overall print speed, not % reduction. You can always double check using the gcode preview.

Travel speed: 200mm/s (higher than Slic3r where it's at 170)

this one won't affect you too much for now but if you're printing a lot of tiny parts in a row then its best to get the nozzle from point A to point B as quickly as possible without oozing. The printer knows how fast it can move, so setting to 9000% won't really do anything. If I'm printing something overnight I'll slow it way down because the travels are by far the loudest movement in stealth mode.