Feature Req: Raft

[rtgoodwin]

I know it's so old-school to use Rafts, but sometimes you have prints or print beds that just DO NOT stick very well with ABS. (Or objects with small bases where it helps to have a nice structure to help remove it from the bed safely).

I don't think it needs to be complex or overly configurable. In the spirit of Slic3r, I think just a "number of layers" field in Advanced would be fine, and let Slic3r handle the flow and pattern.

[Pointedstick]

Ugh, rafts… :)

[rtgoodwin]

I know, I know, but for ABS and certain HBPs/surfaces, it's a good last minute "save" if you're not getting good adhesion. It can be a hidden option with a secret password to open it, if we want to bury it :)

[koolatron]

Raft and support are pretty closely aligned features -- there's a reason they're controlled by the same logic in Skeinforge. I think once support's worked out, raft should be easy.

[IdolCrasher]

No rafts -> ABS users freightened away -> Less Users Overall Users -> Less use/support/donations for Slic3r

That is a terrible shame. It is a very nice slicing program!

[electroleon]

Tried a workaround for this and so far all my small parts print out fine after cleaning off the excess. Use Skirt at a distance of 0 and use a couple of loops (5+) it gives a bigger contact area and sticks a lot better.

[IdolCrasher]

Haha,

Yeah, I have doing that as well. It works like a charm ;)

Sent from my iPad

On Jun 5, 2012, at 7:13 AM, electroleon reply@reply.github.com wrote:

Tried a workaround for this and so far all my small parts print out fine after cleaning off the excess. Use Skirt at a distance of 0 and use a couple of loops (5+) it gives a bigger contact area and sticks a lot better.

---

Reply to this email directly or view it on GitHub: https://github.com/alexrj/Slic3r/issues/123#issuecomment-6111199

[IvanBortolin]

Hi, i'm Ivan Bortolin. I think necessary raft for good print with ABS.

[IvanBortolin]

Hi, for big stamp (es 15x15) raft is necessary because abs undergoes deformation.

[jfpion]

there is the brim function to avoid wrapping it work well

but it could work better if it is possisible to specify the number of layer for the brim. I discover that the diferent paths are not well connected so sometime it is not enough to avoid warping to have the possibility to specify 2 or 3 layer with orthogonal filling to insure strength could be great at least 2 layer with orthogonal filling

I have to do some squares 0.5 mm heigh in the stl to be able to print correctly big parts in ABS

thank you

[Zemciko]

Altough all ppl are mentioning RAFT for ABS i myself realy need it for new batch of PLA.... (ABS stick wo problem for me) On heated bed with hardned glass it realy helps for objects with bigger flat base, preventing bending. I thing same procedure as in skeinforge will be best. (Massive material base and then thin interface).

[IdolCrasher]

Yeah, I think a raft is a no brainer. Of course it is needed.

Pro 3D printers (stratasys) use a raft by default.

Sent from my iPad

On Sep 9, 2012, at 4:22, Zemciko notifications@github.com wrote:

Altough all ppl are mentioning RAFT for ABS i myself realy need it for new batch of PLA.... (ABS stick wo problem for me) On heated bed with hardned glass it realy helps for objects with bigger flat base, preventing bending. I thing same procedure as in skeinforge will be best. (Massive material base and then thin interface).

— Reply to this email directly or view it on GitHub.

[jfpion]

"i myself realy need it for new batch of PLA.... (ABS stick wo problem for me) "

if the abs stick well on you hot bed the pla will stick better ! my experience is you don't need raft on glass and spécially with hot bed and kapton

the only issue is the raft allow you not to tune correctly you bed

but it is a pain in the neck to clean it after

so i print without bed and without pb

[IdolCrasher]

To me, 3D printing with a hobby printer is akin to an art form (like clay sculpting).

In fact, silly as it sounds, I find very little of the process to be "automatic."

It is more like a manual process that takes a lot of time to see the results of your actions! ;)

Just like any artist I would like access to the tools of that trade.

Not every painting uses the largest or smallest brushes; Not every sculpture uses the same set of chisels.

A raft is simply a tool a hobby printer enthusiast might employ in his project.

Raft use is common in "professional grade printers," and should be pretty easy to add as an option.

I assume a raft has not been added yet, because there are more important things to work on.

Never implementing a raft in the program would seem like obtuse thinking to me.

[3space123]

I am using Ultimaker 3mm PLA with a heated bed.

Problems:

1. Every time I print something large, the edges curl up.
2. Whereas the printer used to print pretty well with c. March 2012 version of Slic3r, I now get dozens of nubs all over my print.
3. I get some small burned blobs on the large flat surfaces
4. I am printing a very small piece, 4mm tall, 8mm wide, 1mm thick, convex shape with c. 0.5" radius arc. Instead of printing a smooth arc, Slic3r starts, prints 2mm, moves over 5mm, prints another 1mm, moves back over, etc, until it has done 5 partial arcs, leaving nub mentioned at #2 at each location. The result is hell.

If anybody can advise on any of these, I would really appreciate it.

[IdolCrasher]

Hi,

The nubs are a problem with slic3r 9.7

I hope it is fixed soon.

[3space123]

Thanks IdolCrasher. Any idea about my #4 problem above?

[IdolCrasher]

Try slic3r 9.1 and see if it does not create a better(working) solution for that arc :)

[3space123]

OK, thank you, will do! On a related note, I read that Marlin was much better than Sprinter for things like arcs. Do you have experience using Marlin?

[IdolCrasher]

I am familiar with using and setting up Marlin.

I have never actually enabled the genuine arcing feature within Slic3r though

---

From: 3space123 notifications@github.com To: alexrj/Slic3r Slic3r@noreply.github.com Sent: Monday, December 24, 2012 2:38 PM Subject: Re: [Slic3r] Feature Req: Raft (#123)

OK, thank you, will do! On a related note, I read that Marlin was much better than Sprinter for things like arcs. Do you have experience using Marlin? — Reply to this email directly or view it on GitHub.

[3space123]

Didn't know there was one. Would that help with sprinter f/w?

[IdolCrasher]

I do not believe that sprinter can use arcs that are written into g-cdoe.

marlin can, and slic3r can generate them.  I believe it is still experimental though; I would not hold my breath on it being useful... but who knows.

Maybe someone else has experimented with it.

---

From: 3space123 notifications@github.com To: alexrj/Slic3r Slic3r@noreply.github.com Sent: Monday, December 24, 2012 3:13 PM Subject: Re: [Slic3r] Feature Req: Raft (#123)

Didn't know there was one. Would that help with sprinter f/w? — Reply to this email directly or view it on GitHub.

[alranel]

Let's keep this one about raft, thank you!

[jrherita]

Would love a raft.. please :)

[IdolCrasher]

I second the motion for a raft :)

Sent from my iPad

On Dec 26, 2012, at 22:38, jrherita notifications@github.com wrote:

Would love a raft.. please :)

— Reply to this email directly or view it on GitHub.

[alranel]

Okay, let's discuss what settings need to be implemented for this. The least number, please.

I never printed with raft, so I have no idea about the needs. What about doing always two layers, one being fatter and sparse, and the second one (cross-hatched) with thinner and closer extrudates. Does this make sense? What needs to be configurable here?

[IdolCrasher]

That is fantastic news that you are interested I moving into raft territory :)

I think that very few things should need to be set by the user.

I found a write up: "The Raft is pretty important to a print job. Smaller prints might squeak by without a raft at all, and middle size ones won’t curl up too bad if the raft isn’t there or is defective, but for big jobs, the raft is mandatory, and a build won’t make it if its raft isn’t laid down right.

In skeinforge terminology, the Raft consists of two parts: the Base and the Interface. The base is a set of thick, heavy lines whose purpose is to glue the raft to the build base. The Interface is the relatively thin lines that criscross on top of it to provide a level surface for the build."

[IdolCrasher]

Maybe just give the bottom layer of the raft the width:thickness ratio of that normally goes to the first layer, then let the user set the extrusion percentage of the interface layer? This way the user can dial in very thin interface lines allowing for easy removal of the raft?

[nophead]

Actually any support should have a solid raft at the top because supporting the base is no different to supporting any other horizontal surface. The interface layer needs to be flat and slightly lower as described here: http://hydraraptor.blogspot.co.uk/2012/08/peel-able-support.html

If you watch videos of the UP printer you see that is the technique it uses.

On 1 January 2013 22:54, Adam Herring notifications@github.com wrote:

That is fantastic news that you are interested I moving into raft territory :)

I think that very few things should need to be set by the user.

I found a write up: "The Raft is pretty important to a print job. Smaller prints might squeak by without a raft at all, and middle size ones won’t curl up too bad if the raft isn’t there or is defective, but for big jobs, the raft is mandatory, and a build won’t make it if its raft isn’t laid down right.

In skeinforge terminology, the Raft consists of two parts: the Base and the Interface. The base is a set of thick, heavy lines whose purpose is to glue the raft to the build base. The Interface is the relatively thin lines that criscross on top of it to provide a level surface for the build."

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-11796123.

[IdolCrasher]

Good Point,

I do remember you writing about raft implementation on your blog. I would love to see those idea implemented in Slic3r

[alranel]

@nophead, good thought. As you know, I already implemented your idea in support material. However, I haven't tried using it in combination with a 100% solid interface layer.

What settings do you think we should expose to the end user to customize the raft behavior? Do you think two total layers (base + interface) would suffice?

[nophead]

I used three layers when I did them many years ago before I had a heated bed: http://hydraraptor.blogspot.co.uk/2009/07/thoughts-on-rafts.html

IIRC UP may use even more.

On 2 January 2013 10:08, Alessandro Ranellucci notifications@github.comwrote:

@nophead https://github.com/nophead, good thought. As you know, I already implemented your idea in support material. However, I haven't tried using it in combination with a 100% solid interface layer.

What settings do you think we should expose to the end user to customize the raft behavior? Do you think two total layers (base + interface) would suffice?

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-11803794.

[nophead]

With a heated and level bed I could probably do it with 2 but part of the use case for rafts is without those two conditions.

On 2 January 2013 12:19, nop head nop.head@gmail.com wrote:

I used three layers when I did them many years ago before I had a heated bed: http://hydraraptor.blogspot.co.uk/2009/07/thoughts-on-rafts.html

IIRC UP may use even more.

On 2 January 2013 10:08, Alessandro Ranellucci notifications@github.comwrote:

@nophead https://github.com/nophead, good thought. As you know, I already implemented your idea in support material. However, I haven't tried using it in combination with a 100% solid interface layer.

What settings do you think we should expose to the end user to customize the raft behavior? Do you think two total layers (base + interface) would suffice?

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-11803794.

[alranel]

Well, I can't implement anything without understanding the exact usage cases.

Some people say it's for working with unleveled print surfaces. And I say: that's what the first_layer_height setting for.

Other people say it's for getting better adhesion. Well, Slic3r's got a unique feature called brim. It adds a solid base flange around all parts of the object touching the bed, causing a 100% solid contact area between object and bed. I don't see how a sparse raft would be better than 100% contact. Still nobody was able to answer this simple question: why would a raft improve bed adhesion over using brim?

Some people talk about a sparse, cross-hatched, raft. Other people, including @nophead, talk about a raft having a solid top.

Okay, it's too early for implementing this...

[jfpion]

I agree, I can't see a use for it (except to waste plastic ..)

I think for now there is need for time to solve the dingleberries pb and the start of the perimerters from inside

(and eventually the more external perimeter 1/2 or 13 or 1/4 high to increase the final aspect and keeping a reasonnable printing time)

[alranel]

Let's stay in topic please.

[nophead]

I think I have only use one once since I started using a heated bed 3 years ago. That was for something with a very small contact area at the bottom. I expect brim would handle that.

A raft is pretty much essential if you want to print ABS without a heated bed.

On 13 January 2013 14:45, jfpion notifications@github.com wrote:

I agree, I can't see a use for it (except to waste plastic ..)

I think for now there is need for time to solve the dingleberries pb and the start of the perimerters from inside

(and eventually the more external perimeter 1/2 or 13 or 1/4 high to increase the final aspect and keeping a reasonnable printing time)

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-12194481.

[alranel]

@nophead, how would raft be better than brim even without heated bed? That's something I still don't understand.

[jrherita]

Hi there,

I'm still a learner here, but I have noticed that on some glass surfaces I have real problems getting completely solid layers to stick -- i.e. if you have a large base unit with some holes in it, the outlines all stick fine, and for a little while the first layer fill sticks.. but over time the fill starts pulling itself up.. I was hoping a raft option (with less than 100% fill) might allow for these prints to work...

(I'm printing PLA, tried various temperatures, various cleaning/chemical solutions, etc.. it seems to be something with the glass)

Thanks, John H

On Sun, Jan 13, 2013 at 9:13 AM, Alessandro Ranellucci < notifications@github.com> wrote:

Well, I can't implement anything without understanding the exact usage cases.

Some people say it's for working with unleveled print surfaces. And I say: that's what the first_layer_height setting for.

Other people say it's for getting better adhesion. Well, Slic3r's got a unique feature called brim. It adds a solid base flange around all parts of the object touching the bed, causing a 100% solid contact area between object and bed. I don't see how a sparse raft would better than 100% contact. Still nobody was able to answer to this simple question: why would a raft improve bed adhesion over using brim?

Some people talk about a sparse, cross-hatched, raft. Other people, including @nophead https://github.com/nophead, talk about a raft having a solid top.

Okay, it's too early for implementing this...

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-12194141.

[nophead]

ABS doesn't stick to much cold. I used to use foam board for a bed material. If I built a solid object on it I would not be able to remove it so I used a raft with sparse but strong layers at the bottom graduating to solid at the top. Over time the surface of the board would become cratered so then the raft also was needed to level it off. See http://hydraraptor.blogspot.co.uk/2008/05/stepping-up-production.html

On 13 January 2013 15:18, Alessandro Ranellucci notifications@github.comwrote:

@nophead https://github.com/nophead, how would raft be better than brim even without heated bed? That's something I still don't understand.

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-12194888.

[IdolCrasher]

A better question might be: "Why do professional grade (Stratasys) FDM 3D printers use rafts?"

Sent from my iPad

On Jan 13, 2013, at 6:13, Alessandro Ranellucci notifications@github.com wrote:

Well, I can't implement anything without understanding the exact usage cases.

Some people say it's for working with unleveled print surfaces. And I say: that's what the first_layer_height setting for.

Other people say it's for getting better adhesion. Well, Slic3r's got a unique feature called brim. It adds a solid base flange around all parts of the object touching the bed, causing a 100% solid contact area between object and bed. I don't see how a sparse raft would better than 100% contact. Still nobody was able to answer to this simple question: why would a raft improve bed adhesion over using brim?

Some people talk about a sparse, cross-hatched, raft. Other people, including @nophead, talk about a raft having a solid top.

Okay, it's too early for implementing this...

— Reply to this email directly or view it on GitHub.

[thecrazy]

Id say cause it works but ultimately cause it compensates for unleveled beds thus saving them a bunch of support calls which equals money in their pocket just like the extra money they earn from all the plastic wasted in rafts. There are other ways to do things but it just works for them in many ways.

[IdolCrasher]

Right now, we (slic3r) uses an adjustable thicker bottom layer to adjust for the undeveloped bed. I suggest that a thicker bottom layer is not desirable in all cases and can actually be detrimental to some prints (ie gear teeth)

In fact, a "squished" bottom layer can be detrimental for all kinds of prints.

A raft allows for non-level bed compensation without "squishing" the first layers

[duttonw]

Raft would be excellent as it helps those people who don't want to waste time in getting the bed super level and at the right height such as myself.

Since I changed to abs from pla I now have prints which just won't stick or prints which are so stuck onto the kapton tape over glass that it won't pop off even if placed into the freezer.

This might also come under the process of getting the bed at the right height.

Brim is useful but it's a pain to remove and has the same problems on small parts.

Regards

William d.

On 14/01/2013, at 3:42 PM, Adam Herring notifications@github.com wrote:

Right now, we (slic3r) uses an adjustable thicker bottom layer to adjust for the undeveloped bed. I suggest that a thicker bottom layer is not desirable in all cases and can actually be detrimental to some prints (ie gear teeth)

In fact, a "squished" bottom layer can be detrimental for all kinds of prints.

A raft allows for non-level bed compensation without "squishing" the first layers

— Reply to this email directly or view it on GitHubhttps://github.com/alexrj/Slic3r/issues/123#issuecomment-12207421.

[drk222]

"Other people say it's for getting better adhesion. Well, Slic3r's got a unique feature called brim. It adds a solid base flange around all parts of the object touching the bed, causing a 100% solid contact area between object and bed. I don't see how a sparse raft would be better than 100% contact. Still nobody was able to answer this simple question: why would a raft improve bed adhesion over using brim?"

I tried many materials, some of them curl, especially when you print large solid bottom. Using raft on skeinforge 10mm wider then bottom realy helps. Not only you have bigger area to stick, raft is not printed as filled plane, relativly large gaps helps prevent peel off during material shrinking... And it helps to prevent of curling small details like teeth on gear in first layer.

"Some people talk about a sparse, cross-hatched, raft. Other people, including @nophead, talk about a raft having a solid top."

I think best way is same raft as integrated in skeinforge.

1. thick layer helping to stick and removing bed bumpiness
2. soft layer, for precise print and easy raft removal. if someone need more solid raft, then u can simply change gap between raft lines...

"Okay, it's too early for implementing this..."

Early? maybe too late.

And remeber on 2 head machines raft can be printed from different material in same way as on profi machines, and get nice bottom wo aditional finishing.

[thecrazy]

@drk222 Personally I don't know if a raft is better than a brim for adhesion. We (the community) might just have to agree to disagree. It does seem like there may be specific situations were a raft would be better suited to get the desired results. How to implement it tough... i'll leave that discussion to Alex, nophead, Richrap and the other kingpings of the community.

[IdolCrasher]

Slic3r not having a raft feature is like a car windshield without wiper fluid. There are other ways to clean your windshield, but sometimes you just wish you had some damn wiper fluid.

"Too soon? More like too late" +1 :)

[simonkuehling]

Hi all,

i'd also like to contribute my experience to this discussion. I definitely vote for a raft implementation, because of the following reasons and use cases:

The brim feature is great and often useful, especially for big objects with some fine "arms" on the outline, or for slim walls that span across the build plate. But it is a real pain to remove the brim from all objects with a lot of thin legs on the bottom or - as already mentioned - gear-type outlines that may tend to curl up ocassionally on their teeth. A raft is easy to remove while providing great adhesion for fine details.

Another case that i found just an hour ago: i am experimenting with two nozzles of different diameter (for perimeters/infill) with the bigger one beeing either 0.75 or 1.2mm and notice that it is almost impossible to get sharp corners of the bottom perimeters to stick to whatever surface i try - they peel off as soon as the print head is heading away from the corner. When i build a raft before (manually coded) i can easily print complex outlines on top with those huge nozzle diameters. Oh, maybe it is worth mentioning that i talk about ABS here

My proposal for the user-side implementation is:

• First layer extrusion spacing (distance between extruded, fat lines; from 0 to maybe extrusion width?)
• Interface layer count (one or more layers to "bridge" over the first until a smooth, even surface is present)

First layer extrusion width should be automatically calculated by layer height and nozzle diameter - no need to tweak this as an end user, i think.

[JBFromOZ]

I think a raft for adhesion is an improvement over brim as per this example http://www.flickr.com/photos/87180506@N08/8178701082/in/set-72157631960866866 this uses brim 10mm, on extreme brim the brim extends well outside object area, but still only a very small contact patch. so in preference, a fill of pre-settable fill density that fills the gap between the brim and the skirt, (ie 0.1 density hex) so that an object like this with only a few contact points has a broad base to avoid delamination during printing.

[simonkuehling]

As an update to my last comment regarding the implementation:

• the raft size could possibly be derived from the skirt settings, since it necessarily has to be a little bigger than the skirt outline. the offset could be hard coded as long as there is no wide experience yet.

[JBFromOZ]

I can think of 3 reasons someone would want a raft. • small footprint • warping at edges of flat base objects. • non level bed, so prepare a flat surface for subsequent print.

I believe it is consistent that we could fill the area between brim and skirt for adhesion purposes, and if the "raft" needed to be wider, the skirt offset could just be increased using the existing setting.

Additional parameters might be: • fill density - low density to save on material, and widen the secured base, when compared to solid brim perimeters, • fill type - existing fills eg linear, hex plus a "raft" fill with first layer fatter lines with over-extrude, and next layer a thinner under extruded perpendicular fill or whatever algorithms have proven suitable. • additionally could be set to same number of layers as skirt, or a separate parameter for number of layers of raft fill e.g. so linear fill 2 passes, or "raft fill"

If the print was flat on the base, it would be possible to increase skirt offset around the extent of an object using the existing skirt offset factor.

This would increase the "area of raft" extending past the edge and assist with securing from warping.

To secure an object, as per my example above that does not have much actual contact with the first layer, the brim is a small area even with large brim settings, and in many cases irregular and not even under the object, and a large area exists between brim and skirt.

This filled with a low density wide extent fill should add adhesion under the model extents.

Combinations of these outcomes would be adjusted using skirt offset, raft fill density, number if raft layers, and number of brim perimeters

I look forward to the outcome of these discussion!