Getting tried of layers cracking

[Cymon]

I've built an enclosure, I've printed hotter, I've lowered my filament diameter, I've done everything I can think of but no matter what I do the result is still the same. If I'm printing something with long walls the layers crack. Is there any hope of me fixing this problem. I'm using ABS, so maybe I just need to take the plunge and switch to PLA. Is there anything this will fix this?

[Joseph Chiu]

Are you printing relatively hollow infill?  Maybe thicken the interior walls a bit by cranking up extra shells?

On Wed, Jan 9, 2013 at 8:13 AM, Cymon <joeal...@gmail.com> wrote:

I've built an enclosure, I've printed hotter, I've lowered my filament diameter, I've done everything I can think of but no matter what I do the result is still the same. If I'm printing something with long walls the layers crack. Is there any hope of me fixing this problem. I'm using ABS, so maybe I just need to take the plunge and switch to PLA. Is there anything this will fix this?

--
 
 

[Jason]

Unfortunately, I had this problem, and still do. I spent several weeks discussing it here on the Google Group and really didn't get it completely solved.

I've added a hood and plexiglass windows to the bot, but that didn't help much. However, if I move the bot to a large room with little activity, it works better. But that isn't an option for me most of the time.

The absolute most helpful trick that worked for me is to lower the filament diameter in Print-O-Matic for every print. I lower it down to 1.78mm, and that forces each individual layer to stick together stronger. It definitely helps, but I still can't print medium to large objects like the one you showed.

You could always go the route that Kris Kortwright took, and completely smother your bot in foam: https://plus.google.com/photos/117228107982719398401/albums/5814466883075132385/5814466887054197522

If I understand it correctly, PLA should experience less cracking and warping, but you'd have to tweak around the settings a bit to make it work. 

[Bill Culverhouse]

I've printed a good bit in ABS and PLA, and I would say switching to PLA
would solve the splitting shown. PLA will still "lift" at corners, especially
on large solid/high density infill items, resulting in curved bottomed prints.

But I don't get splitting like this with PLA. I used to see it all the time on ABS prints.

Printing hollow/very low infill might help if you still must print with ABS.

b

--
 
 

[Jetguy]

Well it's basic physics. ABS expands quite a bit when heated to 200C
or higher. Then you lay it down in a layer and it cools to room temp
shrinking quite a bit. Now, you print a layer on top and it does the
same thing, thus trying to break the bond between the two layers, an/
or curl up the bottom layer at the ends. It's built in stress for each
layer and the longer the thread or part, the more shrinkage and stress
built in. Keeping a heated chamber is a false sense of hope. So,
rather than letting it cool to room temp, it's maybe 80C. Well 200+C
to 80C is still over a 100C temp swing and certainly enough to build
the same interlayer stress. SO when the part cools to room temp
(because we cannot keep it in a heated box), you will actually hear
the cracks forming from across the room.

This same effect happend even in injection molded parts. Why do you
think the Z arms on the Replicators warped from HBP heat? The answer
is the injection molding process left built in stress in the part and
when given an outside force such as a temperature swing, the plastic
attempt to releave the stress to bend into a new shape.

You either have to build with a material(PLA) that has less thermal
stress built in, or temper each build in a heated chamber over a long
period of time. The temps required to do that would be near 200C, thus
why the chamber on the bot can't do anything usefull for you. It might
prevent cold draft related stress issues, but won't fix the "built in"
stress of ABS. Plus, as I stated, the extra heat may cause molded
parts of the bot to crack or warp which is a very bad idea. They too
have built in stress and we don't need to find out how much.

On Jan 9, 11:51 am, Bill Culverhouse <bill.culverho...@gmail.com>
wrote:

> I've printed a good bit in ABS and PLA, and I would say switching to PLA
> would solve the splitting shown. PLA will still "lift" at corners,
> especially
> on large solid/high density infill items, resulting in curved bottomed
> prints.
>
> But I don't get splitting like this with PLA. I used to see it all the time
> on ABS prints.
>
> Printing hollow/very low infill might help if you still must print with ABS.
>
> b
>
>
>
> On Wed, Jan 9, 2013 at 11:36 AM, Jason <zen.w...@gmail.com> wrote:
> > Unfortunately, I had this problem, and still do. I spent several weeks
> > discussing it here on the Google Group and really didn't get it completely
> > solved.
>
> > I've added a hood and plexiglass windows to the bot, but that didn't help
> > much. However, if I move the bot to a large room with little activity, it
> > works better. But that isn't an option for me most of the time.
>

> > The absolute most helpful trick that worked for me is to l*ower the
> > filament diameter* in Print-O-Matic for every print. I lower it down to

> > 1.78mm, and that forces each individual layer to stick together stronger.
> > It definitely helps, but I still can't print medium to large objects like
> > the one you showed.
>
> > You could always go the route that Kris Kortwright took, and completely
> > smother your bot in foam:

> >https://plus.google.com/photos/117228107982719398401/albums/581446688...

[Adan Akerman]

Jetguy,

That explanation makes a lot of sense. What is it about the Stratasys machines, then, that lets them print stuff without cracking, if it's not just the 70°C (reported) chamber?

When I'm imagining the printing process, I'm thinking that, though the noodle comes out at 220 (or whatever), it very quickly cools to the temperature of the material on which its been deposited. So although the top layer is, on average, a little hotter than the layer below, it isn't actually able to get completely deposited at that extremely high temperature. So we're not really talking about the thought experiment in which there's an isostrain condition between these extremely different layers:

(activelayer: temp = 220, length = X)

(midlayer: temp = 20+, length = X - 200ish degree shrink amount)

(base layer: temp = 20, length = X - 200 degree shrink amount)

Rather, with the effects of the HBP and the previously mentioned cooling on top, isn't it more moderate than that? And maybe that's why the 70 degree chamber does help?

Thanks for your help thinking through this, I'm hoping to be fussing with elevated temps around the build later today and it would be cool (ha) to cut my bad logic off at the pass.

Adan

--

[Cymon]

I didn't put my printing specs here, so here's some more information:

Problem prints: http://www.thingiverse.com/thing:41154 http://www.thingiverse.com/thing:38194

Usual print settings:

ABS

Layer height: 0.27

Extrude temp: 250c

Extra Shells: 2

Infill: 10%

Extrude Speed: 100mm/s (Sailfish)

Speed Plugin: at least 5 sec per layer

Is it possible I'm printing too hot? Too fast? Too many Shells? Not Enough?

What setting should I fiddle with first?

[Dan Newman]

On 9 Jan 2013 , at 1:05 PM, Cymon wrote:

> I didn't put my printing specs here, so here's some more information:
> Problem
> prints: http://www.thingiverse.com/thing:41154 http://www.thingiverse.com/thing:38194
> Usual print settings:
> ABS
> Layer height: 0.27
> Extrude temp: 250c
> Extra Shells: 2
> Infill: 10%
> Extrude Speed: 100mm/s (Sailfish)
> Speed Plugin: at least 5 sec per layer
>
> Is it possible I'm printing too hot? Too fast? Too many Shells? Not Enough?
>
> What setting should I fiddle with first?

Infill. Kills the print time to increase it, but the three comparisons I'd try are

1. shells=2, infill=25%
2. shells=3, infill=10%
3. shells=3, infill=25%

Mind you, I suspect that 20% is a sufficient change, but I'm suggesting 25% to
hopefully ensure that the effect, if any, is more noticeable.

Usually, when I find that I've had some unexpected layer adhesion issues, it's
because I forgot and used a low infill setting which I had previously used to
just quickly print something out. Other than testing, I try to not go below
15%, but I cannot justify that other than it's a rule which has worked well
for me. (Which is to say, it's a bit arbitrary and I could have easily have
adopted 20%.)

Dan

[Dan Newman]

On 9 Jan 2013 , at 1:18 PM, John Driggers wrote:

> So do you have a rule of thumb as to what size or density object you can
> print in ABS successfully?

Not too sure who you're asking, but as to the density and size folks have
printed to the maximal size of their build volumes and I know from discussions
with MBI that that is a routine test they do. (They did it on the Jetty Firmware
and I believe Sailfish as well.) As to the density, I suspect that
you can run the full gamut from 0% filled to 100% filled. On my ToM, I printed
max. volume and 0% and then again at 50%. First one, 0% fill, just
to "quickly" ensure that *nothing* would get caught or hung up throughout such a
large build. Of course, that was an awefully long span to bridge at the top, but
I wasn't going for a pretty print. And then, I repeated at 50% fill
running many hours and from a *large* spool of ABS. Was part of ensuring that the
3.5 Jetty Firmware worked well and I wanted to do the test before MBI did: better
Jetty and I find any problems than someone else. (And I'm truly surprised that I
didn't burn out my HBP harness in the process.)

> Do you have any references describing the
> thermal coffecient of expansion for ABS?

Try

http://www.engineeringtoolbox.com/linear-expansion-coefficients-d_95.html

However, there are hundreds if not thousands of flavours of ABS. Engineered
for different flow properties, for different mold release properties, for
different strengths, for different Tg's, for different combustion points,
for different densities, for different surface hardnesses, etc. So, any
random coefficient of thermal expansion you find for ABS is just giving you
the order of magnitude. If it's important, you get a much better number from
the folks supplying the ABS to you.

Dan

[Cymon]

Thing is these walls are so thin I'm not convinced that with 3 shells I'll have any room left for infill. In fact I'm not sure that 2 shells does any infill. I may be wrong. I'll give it a try with some storage bins. I can always use more storage bins.

[Bottleworks]

I'm no expert, but 250C is insane in my book.  I print only using ABS and I normally run at 120mm/s @ 230C.  I do find that having a "heated build chamber" helps.  Allow the HBP to preheat for 10 minutes before printing.  That gets some heat in the air.  I also have the advantage of the aluminum arms.  They act as heatsinks and produce a better heated chamber.  

[Eric Cheung]

The physics explanation by Jetguy is correct.

The cracking (delamination) you're experiencing can be attributed to two causes:

1) poor layer-to-layer adhesion

2) thermal stress (often simply referred as warping)

To solve 1, you generally increase temperature. To solve 2, you generally decrease temperature. Ouch! 

From a strength perspective, more continuous material (shells) build up more thermal forces. But more material gives you more load capability. Ideally, I would want to print [my parts] with flimsy thin long walls, and super bulky strong corners. Though I'm not sure if RepG have options like that.

Though I have a rep2 with PLA and I don't have as bad a problem, I'm embarking on a journey to print Nylon, and the issue is quite a bit worse than what you have shown. However, the underlying physics are the same. Hotter increases layer-to-layer strength, but increases thermal stress at the same time.

However, it looks like your base is flat, which is what I cannot achieve at this point ^^!

[Mark Durbin]

Hi Cymon,

I'd have added circular lugs at the corners to pin it to the bed, but I always had problems building with ABS.

I did use to stop the print once I had the lugs printed and put bulldog clips on them to hold down the print until the build completed.

But I would then get splitting on anything under 3mm.

PLA does not give me these problems fan cooled as it leaves the extruder, it solidifies very quickly and doesn't shrink much.

I regularly print items 200x120x120 with flat bottoms and tall thin sides (1.5 mm).  You still need to get it well seated on the platform, but after that it's a breeze.

I have found that printing on Slightly roughened Kapton tape gives me a better base finish and tighter control over the base spread than painters tape, but plenty of people swear by it.

--
 
 

[John Driggers]

Specifically Jetguy Dan, but I appreciate your answer, as it's much more in line with my personal experience than Jet's blanket statement that basically says it's impossible to get good prints out of ABS due to thermal expansion.  I also concur that just about ever filament I've used exhibits a different expansion rate, and based on the melt and glass transition temperatures, I suspect a pretty decent difference in the A / B / S ratios.

Now, get me some sailfish goodness on some other hardware - and in a perfect world, support delta motion! :-)

[bagelturf]

One thing that has worked for me is to add a thin (1mm or less) wall around your model and make it hug the outline really closely, not more than a couple of mm away. The wall helps keep the heat in while the model is building so it is build with less stress.

[Andrew Russell]

Cool idea bagelturf!

I think this is what the skirt plugin in SF is built to do.  I've only used it to keep the anchor line away from the first layer of my prints, but I've thought about using it to insulate the print. It'll be a thinner wall than 1mm since its a single extrusion line.  I have my gap over perimeter ratio @ 2.  This gets pretty close to the print but doesn't touch it.  

I wonder if this would help the folks working with taulman filament, or if the upper skirt layers would eventually warp and collide with the hot end.

-Andy

--
 
 

[Cymon]

Joseph Chiu's recomendation on day 1 fixed it, I just haven't had the time to test until yesterday. 3 shells, 25% infill.

The problem seems to be that high angles on flat walls are high stress points for thermal shrinkage. In fact in the print above there is still a crack in the corner nearest the camera where the 3 shells was the only thing between the wall and the fill area. (But the narrowest point on the other side was solid.) I've seen this before but only on tall and wide flat walls with 90 degree angles like the back of the iThrone. Where the model had a nice gradual curve doesn't seem to have this problem. This would suggest to me not to design 90 degree angles on walls of more than 30mm that are taller than 20mm if you can help it for maximum print ability in the future. I wonder if this sort of thing could be identified in the slicer and compensated for by building a thicker solid wall just in the corners. Kind of like rounding off the inside corner and how much rounding would be necessary. The problem is identifiying the minimum necessary to see the problem and then figuring out the minimum necessary to fix it, both of which I suspect are considerably more time and plastic than a 20mm calibration cube.

[Matthew Marlowe]

I've been fighting building large objects too with a replicator 1 and ABS -- 95% of my prints are take 8hrs or less, and if the build plate is leveled enough, preheated for 15 minutes, and the enclosure in place - there isn't a problem.   But, I am struggling to avoid cracks and other deformities in walls for the large prints that can take 12-24hrs to print.

I've identified what I think are the two major culprits:
* lifting corners - we're all familiar with this, the 90 degree corners at the base of a print are the weakest point thermally holding the base down and no matter how well the machine is tuned, at some points these start to lift if just a little bit....the longer the print is, the more stress I think that builds up during a print and the result is that 8-12hrs in, cracks may show up above the corners near the middle or the print.  Replacing 90 degree corners with multiple smaller angles might be a way to decrease cracks, but that would definitely limit designs..
* Any sudden increase in the width of walls seems to increase cracking (e.g. handles/etc)

I printed the following out a few days ago with +2 shells, .22 layer height, 60/60 speed, and 15% fill - the print ran overnight, took 16hrs, and the results were disappointing...I'm still not sure what can be done to print it successfully, other than to increase layer height further, reduce wall thickness, or possibly increase fill (using replicatorg/MBI firmware 7.0). 

http://www.flickr.com/photos/deploylinux/8380029246/in/set-72157632409758987/
http://www.flickr.com/photos/deploylinux/8380032278/in/set-72157632409758987/
http://www.flickr.com/photos/deploylinux/8380033962/in/set-72157632409758987/
http://www.flickr.com/photos/deploylinux/8378957159/in/set-72157632409758987/

[Shawn]

I'm sure you've tried this already because you've said you are doing
overnight prints, but...

You say you have the enclosures. I'm assuming you mean the 2 sides, the
front, and the hood? If so, I've had some luck by letting everything
come to room temperature normally before I remove any part of the
enclosure. So for taller prints, I might leave the completed print in
place for an hour before taking the front panel off. I find that I get
fewer cracks when I do this, more when I remove the part right after
it's done.

Also, you mention the lifting corners. I have found this to be a
symptom of cooler ambient air getting into the chamber, or that I need
to re-level my platform. Funny, these days I find almost EVERYTHING a
symptom that a relevel is needed. lol But, I am also finding that I
only need to level once or twice a month these days.

I have not done 8+ hour prints for a while now - not since Sailfish
acceleration, so I can't really say if I'm dealing with the lift issue
properly on the long long prints. And it sounds like you've addressed
the usual points... Maybe slowly move a candle around the edges of your
enclosure to see if maybe you still have a draft issue? Oh, try raising
or lowering your platform temp too - initially I found I needed 115c to
avoid peeling, these days I'm good with 110.

Best of luck!

> --
>
>

[Matthew Marlowe]

Shawn,

Yes, I have the full enclosure with sides and hood, and the cracking
appears before the print is removed from the plate. However, perhaps
leaving them to wait for 1hr after the print may reduce how much the
cracks eventually widen.

The build plate is leveled enough that most 0.15mm layer prints of
8hrs or less succeed. I've manually leveled the plate via 30 minute
pre-heat, homing, dropping 0.1mm, disabling steppers, moving extruder
all across plate until z offset is as identical as I can make (there
is a slight warp starting in the center and moving back somewhat). I
use skirt on all my prints and watch the skirt to make minor
adjustments so that the skirt comes out mostly flat and even thickness
in each quadrant. I also have the hbplocks.

Printing is done with hbp set to 115. I clean the plate before each print.

Most of my prints come out fine, but all of my long prints of 8hrs+
have at least minor if not major cracking defects.

I've ordered the hbp arms. What I haven't tried yet is hairspray.....
and there are a few small places in the hood where air can go
in/out....I may experiment with covering some of those with tape for
time being.

Matt


--
https://www.twitter.com/deploylinux