Anyone got any tips for printing when the first layer has lots of tiny part?

[tonycstech 5]

you don't need to learn anything the program is almost the same in its settings.lol actually make more sense.

[tonycstech 5]

problem you are having cannot be solved because of the way cura works. Its hopeless, trust me.

I even tried to develop an application that would fix the G code and insert retraction where it needs to be but I couldn't because of the way cura generates the code it's impossible so it's hopeless.

Cura developers have to make slicer changes to fix it and u asked and asked and asked again but all I get in return is "try this or that".

none of the solutions would work because of attraction in cura is distance-based so nothing would solve the problem

[jweaver 2]

problem you are having cannot be solved because of the way cura works. Its hopeless, trust me.

I even tried to develop an application that would fix the G code and insert retraction where it needs to be but I couldn't because of the way cura generates the code it's impossible so it's hopeless.

Cura developers have to make slicer changes to fix it and u asked and asked and asked again but all I get in return is "try this or that".

none of the solutions would work because of attraction in cura is distance-based so nothing would solve the problem

 

I still think I have a bigger problem not caused by CURA.. My hot end seems to become "unprimed" after a retraction and lays 'air' for a few mms.. So currently, this causes shapes to be incomplete... If CURAs logic changed so that it retract at different points, then surely the problem would move to the next shape it creates AFTER retracting where the initial bit of PLA wouldn't come out.

What I really need to happen is that the printer does 2 passes for every shape, to be sure that its complete.

This is of course assuming that I understand what I am talking about and retraction is even my problem.

As I said, my problem seems to be that I extrude a bit of "air" after a retraction and this would happen irrespective of where the retraction occurred.

I am thinking that the better option would be the lay the entire first layer down. With no retraction.. And the clean it up after the print. As I assuming that my problem is all based around retraction.

Tonight I am going to do some tests by turning retraction off and seeing what effect this has on the first layer.. Perhaps then I will actually understand what the issue is.

Jon

 

[valcrow 144]

This sounds like a printer problem and not a cura problem to me. Mine doesn't seem to have any lag time laying down filament caused by retraction.

I printed this chain

it is 1mm thick, each chain link is individual on the first layer so I imagine it is similar to what you're trying to print. Tiny tiny pieces with retraction on each link. They are only connected half way through the link. It was printed successfully on Cura with tape and on glass with UMO and UM2. Needed some alcohol sticking powers for the tape though.

 

[tonycstech 5]

Problem is part to blame the blue tape printing pla.

I don't want to prep the tape for print because it makes it last only one or two prints. If I don't prep the tape, it lasts for many prints and removal is very very easy. Prepping the tape with alcohol is like applying abs juice for printing abs.

Tell me why printing exact same part using crwftware has no problems and printing with cure has problems ? Its not hardware, its software.

I spent days just trying to learn to why models with multiple perimeters fail to print on a blue tape and found that its because there is no retraction taking place before each perimeter.

After I tried craft ware and problem went away,I instantly knew why.

[jweaver 2]

Problem is part to blame the blue tape printing pla.

I don't want to prep the tape for print because it makes it last only one or two prints. If I don't prep the tape, it lasts for many prints and removal is very very easy. Prepping the tape with alcohol is like applying abs juice for printing abs.

Tell me why printing exact same part using crwftware has no problems and printing with cure has problems ? Its not hardware, its software.

I spent days just trying to learn to why models with multiple perimeters fail to print on a blue tape and found that its because there is no retraction taking place before each perimeter.

After I tried craft ware and problem went away,I instantly knew why.

 

Have you got a good example to prove your point with? I am going to have a crack with Craftware tomorrow.

[tonycstech 5]

Sure. I will make a video tonight if I won't forget.

I will print small object with multiple perimeters using cure and craft ware. Cura print will fail at first layer, guaranteed.

[aviphysics 9]

problem you are having cannot be solved because of the way cura works. Its hopeless, trust me.

I even tried to develop an application that would fix the G code and insert retraction where it needs to be but I couldn't because of the way cura generates the code it's impossible so it's hopeless.

Cura developers have to make slicer changes to fix it and u asked and asked and asked again but all I get in return is "try this or that".

none of the solutions would work because of attraction in cura is distance-based so nothing would solve the problem

 

Well, if you were feeling especially ambitious, you could write a program that would create a list of G1 movements in the outerwall sections. Then find all G0 moves which cross any of those G1 paths and turn them into retracts. Computationally, this ought to be entirely possible. Perhaps just a little bit of a PITA. Not at all sure how long it would take. My gut says it would be no more computationally intensive than an n^2 problem, if you didn't try to do anything fancy and just took the list of G0 moves and tested each one against the G1 moves. (I am assuming that the number of G0 moves and number of G1 moves grow proportionally, which they probably don't)

 

[tonycstech 5]

I am and i tried and i failed because there i no reset taking place i found it impossible with my level of programming skill.

Maybe possible mut way too time consuming.

Craftware is already implemented vector slicer so as soon as they release it for windows, i am a permanent gonner.

Not that i think anyone would miss me, but just saying.

I suggested them to implement vector slicer and in just a few days they did because i was not the only one having the issue.

But with CURA, every time i bring up this issue, all i get is "try this or try that" Problem is at the CORE, not with the settings or the plugins.

Seeing how ignored i am around here, i might as well just pack and leave.

 

[jweaver 2]

Sure. I will make a video tonight if I won't forget.

I will print small object with multiple perimeters using cure and craft ware. Cura print will fail at first layer, guaranteed.

 

Video would be really helpful as i would love to see how CURA is working for you and compare it to my problem.. But i was thinking about an STL file..

If you do a vid, perhaps i could try the same STL as you to compare...

With CW. Can you print directly to the printer? Does it support most printer types (for me a Printrbot), or will i just use it to slice and then use something else to print the GCODE

[nallath 1,120]

You are not ignored. Just because we don't implement your function does not mean we ignore it. It simply means we don't have the resources to implement every single feature request (we would need triple the number of software engineers to even get started).

Daid is generally quite blunt and straitforward in his response to feature requests, which i can imagine is ofputting. This is mostly due to dutch culture (eg; honesty / being straightforward is seen as something good) and him being an engineer (which tend to be more direct anyway). Combine that with a very large amount of people writing very poor (no details, no 'why do you want this function', or 'what is your problem in the first place) requests, you get these kind of responses.

I'd also like to add that just because a change removes a problem, it does not always solves the problem. Nor does it mean is is the best way to do it. We're currently trying to fix the big pile of code that our software has become (partially because we kept adding 'fixes' to problems, making it a huge load of "but in this case you need to solve it differently".

 

[tonycstech 5]

Printing exact same STL sliced with CURA and CRAFTWARE with very similar (if not exact) settings.

1.2mm wall (doesnt matter)

0.9mm bottom/top (doesnt matter)

Distance for retraction to occur is the same for both, but Craftware will retract at perimeters as well as when distance says so. Cura would only retract when distance says, but not at when it changes perimeters.

Combing is enabled in both cura craftware. Craftware does not have combing option, it just had it as a standard process, and its not combing as far as cura does. Cura sometimes would make the nozzle travel into so many places before reaching the next print point, nozzle ends up being completely dry (but thats another issue that will be ignored by CURA developers as well as this one)

Video will be available once finished uploading.

http://youtu.be/M0w8KVHPGi0

 

[Daid 306]

Speed A is not Speed B. So your test is flawed right there. Pretty sure Cura won't give the "rip of first layer" problems if you set the speed to the same amount for the first layer, instead of using vastly different parameters.

 

[jweaver 2]

Printing exact same STL sliced with CURA and CRAFTWARE with very similar (if not exact) settings.

1.2mm wall (doesnt matter)

0.9mm bottom/top (doesnt matter)

Distance for retraction to occur is the same for both, but Craftware will retract at perimeters as well as when distance says so. Cura would only retract when distance says, but not at when it changes perimeters.

Combing is enabled in both cura craftware. Craftware does not have combing option, it just had it as a standard process, and its not combing as far as cura does. Cura sometimes would make the nozzle travel into so many places before reaching the next print point, nozzle ends up being completely dry (but thats another issue that will be ignored by CURA developers as well as this one)

Video will be available once finished uploading.

http://youtu.be/M0w8KVHPGi0

 

Thats pretty much exactly what I see. So can I try to understnad something.. Is the blobbing and incomplete shapes caused by Retraction?

If so, what would happen if Retraction is turned off?

The main difference I see is that CW is laying down all the perimeters first for each 'shape'. Which is exactly what I said the problem was above.

It seems that CURA does the first outline, then moves off leaving some "open" (due to the extruder not being primed) and them comes back later ripping them off.. But CW seems to be laying several perimeters, "sealing" them closed and helping them stick.

Is that the behavior you have been trying to describe?. If so, I absolutely see how CW handles this better.. Its the very fact that CURA does each perimeter separately which causes this..

 

[IRobertI 520]

I measured the first perimeter after the skirt to be printed roughly twice as fast using cura in that video. ~18sec for Craftware and ~10 sec for cura. That's quite a difference.

 

[jweaver 2]

I measured the first perimeter after the skirt to be printed roughly twice as fast using cura in that video. ~18sec for Craftware and ~10 sec for cura. That's quite a difference.

 

Lets not get hung up on speed.. I have EXACTLY the same problem as shown here.. And the "speed" doesn't change anything..

The root of the problem is the way that CURA lays down its perimeters, combined with "dry" extruder laying down air for a few mms.. This leads to gaps an blobs which gets snagged.

It seems that Craftware does it differently and adds more perimeters to each shape before moving on..

 

[tonycstech 5]

Problem is not that craftware prints more perimeter then cura before switching, problem is retraction. I dried dissabling combing and it fixed the problem but also created another where cura would retract too much if combing is enabled. Cranking the retraction values down does not help, it will still retract based on the distance and not on the situation.

I can make craftware print 5 times the speed of cura result will be the same.

U think am an idiot?

One answer I didn't expect was yours.

U could say it was because cats have balls, I would find it easier to believe and castrate my cat.

Unfreaking believable. That's exactly y this issue will live on forever in cura, its permanently cured into it.

Here is a video i just make just for you.

Your estimated speed was almost perfect, it was 50% of the 30mm per sec that i wasnt aware of so i cranked it up to 100%, resliced the model and even gave it 250% speed which would make it be 75mm per sec 1st layer. I dont print that fast even on a second layer but for you i did. enjoy.

http://youtu.be/vs4nTeKyvKs

 

[jweaver 2]

Problem is not that craftware prints more perimeter then cura before switching, problem is retraction. I dried dissabling combing and it fixed the problem but also created another where cura would retract too much if combing is enabled. Cranking the retraction values down does not help, it will still retract based on the distance and not on the situation.

I can make craftware print 5 times the speed of cura result will be the same.

U think am an idiot?

One answer I didn't expect was yours.

U could say it was because cats have balls, I would find it easier to believe and castrate my cat.

Unfreaking believable. That's exactly y this issue will live on forever in cura, its permanently cured into it.

Here is a video i just make just for you.

Your estimated speed was almost perfect, it was 50% of the 30mm per sec that i wasnt aware of so i cranked it up to 100%, resliced the model and even gave it 250% speed which would make it be 75mm per sec 1st layer. I dont print that fast even on a second layer but for you i did. enjoy.

http://youtu.be/vs4nTeKyvKs

 

Got to say, I am convinced! This is exactly the problem I see.. And CW will certainly help due to the way that it draws out its first layer.

I just hope they bring out their Vector based version soon, and I will give it a go. I just hope that it prints directly to Printrbots!

 

[tonycstech 5]

I think convincing people that hell is real is much easier then convincing them that cura can't handle first layer.

I wish it was the other way around, maybe then world would be a better place.

[ccw1984 1]

I was having kind of the same issue, where the very first part of the first line put down on the glass would curl up and not stick then cool in that position and get caught on the second pass. I fixed it by using hairspray instead of the glue stick (IMHO it's better because it provides a more even layer and I can ensure all points where it will lay down a first layer will be sticky) and by leveling the bed REALLY well. Now it sticks the first layer to the glass every time with no problems (so far, knock on laywood)

 

[tonycstech 5]

these replies with glue and hairspray and all sorts of WORK-AROUND's :mad:

Why work around ? Why not just fix that problem and avoid any workarounds ?

I mute this post, its pointless.

 

[jweaver 2]

Problem is, i dont grt the feeling this would be easy to fix as it appears to be a problem with the logic that Curas slicer uses.

Since the devs are focussed on more urgent issues i cant see this being looked at, espcially if they do not agree there is even a problem..

[aviphysics 9]

Printing exact same STL sliced with CURA and CRAFTWARE with very similar (if not exact) settings.

1.2mm wall (doesnt matter)

0.9mm bottom/top (doesnt matter)

Distance for retraction to occur is the same for both, but Craftware will retract at perimeters as well as when distance says so. Cura would only retract when distance says, but not at when it changes perimeters.

Combing is enabled in both cura craftware. Craftware does not have combing option, it just had it as a standard process, and its not combing as far as cura does. Cura sometimes would make the nozzle travel into so many places before reaching the next print point, nozzle ends up being completely dry (but thats another issue that will be ignored by CURA developers as well as this one)

Video will be available once finished uploading.

http://youtu.be/M0w8KVHPGi0

 

I have never had that kind of problem with Cura on the Ultimaker Original and I just printed something with a similar sort of first layer. BTW, what was your first layer thickness.

It would be nice if Cura could do a better job of minimizing combs and retractions overall, but that is another story.

 

[Daid 306]

I have never had that kind of problem with Cura on the Ultimaker Original and I just printed something with a similar sort of first layer. BTW, what was your first layer thickness.

 

That most likely due to the UMO having a better nozzle shape, which reduces the chance of the "curling up" effect.

Which brings me to this reply:

 

these replies with glue and hairspray and all sorts of WORK-AROUND's :mad:

Why work around ? Why not just fix that problem and avoid any workarounds ?

 

You're actually using a workaround. The problem you're having is that the material curls a bit up and sticks a bit to your nozzle. Because CW prints the perimeters at in a different order you get more surface area for it to stick on the bed. So you have more grip on the bed because it has double the surface area of the filament on the bed.

However, that IS a WORK-AROUND (feel free to add your angry smiley yourself). As the actual problem is that the material is curling up towards your nozzle and sticking to it.

The whole of 3D printing is full with workarounds. And it's usually just a choice of selecting the most logical workaround for the problem that has the least amount of side-effects.

Most likely your problem is also gone if you set the "Initial layer line width (%)" to 150%. As that would increase the stickiness to the bed.

You can also clearly see in your 2nd movie that the speed has a huge effect on how easy you rip it off. The first line is pulled off for a few cm before it finally luckily sticks.

You're pretty much looking at the tests to see the results you want to see. So no surprises that you come to the same conclusion every single time.

EDIT: As for code changes. This requires quite a few changes in the path order planner. Which isn't as simple as it sounds. As inner wall lines from multiple sections could have been merged into a single section after shrinking. So you get complex logic where a section can only be printed if all touching inner lines have been printed. Quite a bit of complexity for not that much gain.

I'm also unsure what it will do for print quality on higher layers as the cooling will be differently.

 

[DidierKlein 729]

And you can also specify the outer shell speed to be slower than the rest in CURA (advanced tab -> Outer shell speed).