Horrible looking infill on Ultimaker 3

[Dreyfusduke 0]

Hello,

I have a brand new Ultimaker 3 ext and I can't seem to get the infill to work correctly (cura). Using, PLA, everything seems to print well except the infill. It just seems the machine jerks around and half prints a mess instead of a usable infill. I have slowed the entire print down to 30mm/s and temp up to 210c. It doesn't seem to change anything. I have done the "atomic pulls" and the nozzle is completely clean. Funny thing is, I can use Simplify3D and the infill is just fine. Any ideas? Anyone care to share their settings?

[neotko 1,417]

Totally agree, is. Cura issue. Cura standard profiles use the infill without care and leaves all the trash (more or less) without retractions and making a inner structure quite horrible.

I think that can be fixed anyhow, but I never messed around with Cura long enough since S3D works (but is really really slow for dual compared to Cura atm). S3D guys toldme on a mail that they are implementing a heat/cold system like cura for next release, so probably in a few X time you might not need cura at all.

Until then, I would start by unhiding all the many options Cura has and play with retractions distance and combing settings.

[tomnagel 126]

The default print profiles in Cura are made such that 99% of the models print fine. Also the infill should print fine.

What is wrong with your infill? maybe you can illustrate with some pictures?

Are you using Ultimaker filament?

[gr5 2,167]

I changed my settings so that the infill speed was the same as the shell speed and that helped me quite a bit.

[tomnagel 126]

I changed my settings so that the infill speed was the same as the shell speed and that helped me quite a bit.

 

Did you lower the speed of the infill? Or did you increase the speed of the outer shell?

I'm interested in why you find the quality of the infill important. I see it mostly as a support structure for the top layers. (Strength mainly comes from the number of shells). And in that sense, I don't really care how the infill looks like, as long as it forms the support structure that I need. So then, the choice to print it as fast as possible sounds like a good choice to me. Also, the choice to dump ooze into the infill structure sounds smart to me as well, if that (together with some other measures) removes the need for a wipe tower.

But again, I would really like your opinion on this, Neotko and gr5.

[neotko 1,417]

@tomnagel

First is a matter of printing right. I want to have as much control over the extrusion as possible. Making the infill be the trashbox is a bit of a weird fix. BUT, I understand that it helps to keep the print times beyond insane. So... Why I don't like to have a bad infill

a) I do many 3-5mm tall objects, infill must be perfect to be able to do perfect toplayers with just 3 0.2 toplayers, not one more. (more top layers = more time + more filament + more costs)

b) I need to have as much control over when it extrudes/doesn't extrude, without filament loss in the middle. Otherwise I can't print very small details with thousands of retractions.

c) I need constant speeds for the prints, otherwise, I get errors on perimeters, bad toplayers. Every single detail must be controlled when you want to print at 60-70mm/s 0.2 layers at 3000accel/20yerk. Any fail or detail, ruins the surface of how it looks.

Mind that I don't do industrial parts, I do things that look nice and fell nice. I try to make stuff that's nice to have, and non 3D printers can't understand why a top layer has small interlines of holes. Also I have work optimizing every setting and speed of my prints to make decent good looking prints, that also can be printed in 25minutes and don't need more than 1-2m of filament (1.75mm filament).

So for me, every detail is important. Every retraction, movement, speed. Making good looking toplayers isn't easy at all on 3d printers.

Also, infill plays a roll on this, since you never have the same amount of infill area, unless you print just 1 object. So, when the infill area changes, if the extrusion isn't perfect, you can carrie that blob-pressure-changeof-extrusionspeed, to another part of the visible area.

Imo, I prefer, and need, constant speeds with a max fluctuation of 20%, sometimes 30% if necessary.

Edited January 25, 2017 by Guest

[gr5 2,167]

I find at every speed change you get over or underextrusion. So when you go from fast infill to slow shell for example you can get overextrusion for a few millimeters. Until things stabilize again. Creating an annoying bump on the wall somewhere.

[neotko 1,417]

@tomnagel

Another reason to have good infills

https://ultimaker.com/en/community/1467-post-your-latest-print?page=160&sort=#reply-170017

They can be used for the print. Also, imagine a flexible part, where the infill makes parts to be less or more rigid.

IMO every part of a print is important. For structure, quality and precision.

[Dreyfusduke 0]

After the Cura Update to 2.4 I figured I would give Cura another chance......and nope, infill is still junk. No idea how they can even call their "stock" settings remotely good. The outside of the part may be ok to look at, but the inside is just a messy glob of infill failure. You have to slow it down to 25% and have a 2 day build. (and it still looks, and functions, poorly). There are plenty of example why a good infill in needed.

Its like Cura programmers are wanting people to buy Simplify3D. LOL

[neotko 1,417]

I agree. But they want to print faster, so they don't make a good infill to gain speed. IMO is a bad decision. If a profile (or machine) is juat slow, then is just slow. Making infill with bad quality is the only way they can keep a decent ratio of print speed. Because for example um2 head vs um3, since um3 is heavier (and other stuff i leave for a rainy day) they need to use very slow accel/yerk to keep decent quality (and also they use 500 accel/5yerk for thr outside so it looks better than um2). So the only way to print faster is by making low precision infill.

Anyhow. You can unlock the speed hidden settings and set 'manually' all speeds to the same speed. This ofc, since the accel/yerk stills there will make some artifacts from different nozzle size & width for each outter layer, inner, infill). Basically they have make a mix of mm3 playing with all this settings to get better looking prints but as fast as they could. Also they need to fight the bowden pressure changes, making all a big mess that they where able to fix.

Um3 profiles are tailored to give the best possible quality & speed. But anyway, I don't think infill is the place to put all the crap

[yellowshark 153]

Yup, with the exception of the initial layer, I have always printed with the print speeds all the same to ensure the quality of the perimeter wall. If I do have a piece with a filled large x/y geometry then I may push up the infill speed for speed purposes, but if I do I will probably add an extra wall or two to protect the perimeter wall.

IMHO a consistent pressure to protect the perimeter wall is one of the best things you can do if clean surfaces are important to you.

[flowalistik 181]

Please improve how infill works on Cura, it goes too fast, it's too thin. Cura magic is cool, but sometimes it just doesn't work.

It's sad to see my Ultimaker 3 ruin the infill after some beautiful walls.

[ultiarjan 1,223]

And to add, not only is the infill speed set too high imho in UM3 profiles, on top of that also the infill line width is set higher than the print line width, which is a double mistake imo as it not only increases the volume/s the machine needs to push out, it also provides a worse support structure for top layers.

The main issue I have with all this is that it's all in settings that are hidden by default, so many users will have no clue...

This type of default setting should only be in a "Super fast, forget quality" profile, not in a "normal".

[Dim3nsioneer 557]

Maybe it's a silly idea but what about introducing an acceleration/deceleration for the line type change? Such an acceleration would go over a certain print length, i.e. it might include even a large number of print moves. The reference speed should of course be the slower one. If a print is not large enough the infill speed could be reached it would just print the infill at a lower speed.

An alternative could be to print the infill a bit hotter and have the firmware of the printer adjusting speed to current temperature. But that's more tricky as it depends on both the material properties such as density, heat capacity etc. and the heat capacity and heat transfer rate of the hotend.

[Traveler218 1]

And I thought it was something I was doing wrong:O

This was a very educational post, for those of us with limited Cura experiance.

Thanks!

[kmanstudios 1,120]

And I thought it was something I was doing wrong:O

This was a very educational post, for those of us with limited Cura experiance.

Thanks!

Ditto and, I for one, am unanimous in this

[ultiarjan 1,223]

Oh, and one more thing not helping, I noticed again that Cura 2.7 is creating a ridiculous amount of retractions on infill, using a default profile (I only changed first layer height, and infill speed)

[Labern 774]

I changed my settings so that the infill speed was the same as the shell speed and that helped me quite a bit.

I'm interested in why you find the quality of the infill important. I see it mostly as a support structure for the top layers. (Strength mainly comes from the number of shells). And in that sense, I don't really care how the infill looks like, as long as it forms the support structure that I need.  So then, the choice to print it as fast as possible sounds like a good choice to me. Also, the choice to dump ooze into the infill structure sounds smart to me as well, if that (together with some other measures) removes the need for a wipe tower.

One problem with rubbish infill is you get little bits sticking up which then stick to the nozzle. Then not only do you get hard deposits on the nozzle but you can get burnt marks if it falls off. This is not so bad will PLA but shows up in a lot of other materials. But you do get some blobs every now and again from this which can blemish the external shell with PLA.

Also what about transparent filament. I kinda like to see the infill in some transparent prints. It makes a nice design or something to demonstrate what the internals of a print is like. It's hard to promote The UM3 when you show people a print and they ask what all that ugly stuff is on the inside when it's under extruded or full of ooze.

Put it this way. If there was a tick box saying clean infill, even if it took longer to print, I would have it ticked 90% of the time.

Edited August 22, 2017 by Guest

[ultiarjan 1,223]

Think it should be the other way around. Good infill should be the default. A "give me extra speed I don't care about my infill" tickbox would be more logical... But even better just a high speed low quility profile... Its hard enough to work with the ton of tickboxes cura already has.

[kmanstudios 1,120]

Just as an aside, I am using the quick infill to give me a bit of an 'ice' look inside my transparent objects in certain cases. Actually has a nice, crusty, frosty look.

Other times, I have just slowed down the infill to get a cleaner output when visible by design.

[Labern 774]

I wasn't suggesting they add a tick box for clean infill. Just indicating that I would prefer having sturdy infill with no oozing over print speed 90% of the time.

Edited August 22, 2017 by Guest

[Labern 774]

Just as an aside, I am using the quick infill to give me a bit of an 'ice' look inside my transparent objects in certain cases. Actually has a nice, crusty, frosty look.

Other times, I have just slowed down the infill to get a cleaner output when visible by design.

 

Yes this is easy done but you still can't get away from Cura's travel problems unless you turn combing off and have it z hop on retract which isn't the nicest for every case.

[kmanstudios 1,120]

 

Just as an aside, I am using the quick infill to give me a bit of an 'ice' look inside my transparent objects in certain cases. Actually has a nice, crusty, frosty look.

Other times, I have just slowed down the infill to get a cleaner output when visible by design.

 

Yes this is easy done but you still can't get away from Cura's travel problems unless you turn combing off and have it z hop on retract which isn't the nicest for every case.

 

Yeah, that is still looming on my horizon and why I tend to see things a bit different. Not that I disagree, just saying it is not part of my experience yet as I am still learning bunches o' bunches and bunches o' bunches to learn yet.

Still da clumsy noob when ya get down to it.

[cjs 34]

On 25.1.2017 at 10:46 AM, tomnagel said:

But again, I would really like your opinion on this, Neotko and gr5.

After reading all the replies regarding the infill settings are we going to see a change in the Ultimaker stock profiles? 

[kmanstudios 1,120]

Is this me being indiscriminate or is this not decent infill? This is using the basic settings, nothing really changed but the main temp from 200°C default to 195°C. This is 3.2 Beta. And, I have seen much worse infill in the earlier versions of Cura from about a year ago. But, is this not acceptable? The walls are really coming out nicely too. No priming tower or ooze shield.