Why does 618 nylon warp?

[Michael Buffington]

I thought I'd solved the issue of 618 nylon warping while printing by printing onto a Nylatron plate.

Today, however, about halfway through a somewhat large print, it began warping right where I thought it might from having watched previous failures.

It seems as if the warping is the greatest when long lines of thread cool. For example, the object I was printing had no straight lines for the first 10 layers aside from the minimal fill. The longest straight line was probably 1mm. At layer the first of 3 top layers began. Let the long straight lines, and therefore, the warping.

Aside from creating objects that have minimal straight lines, is there any way of avoiding this? Would heating the chamber help?

My current strategy is to try and see if I can trick the slicer into reducing line lengths.

Here's an example - the first screenshot is the object's underside. This prints fine up until it's filled in on the top:

Here's what I've done to try reduce the long lines - added 0.1mm slots:

That gives me slices that look like this:

I'll know within an hour if the strategy works, but in the meantime would love to hear everyone else's thoughts.

[Adan Akerman]

This is a fascinating approach. I'll definitely be watching for your posts to see how it turns out. 

There's a good logic to it, but I wonder what it will do to part strength? 

My initial experiment with chamber heating was a.) extreme (~80C, oops) and b.) ineffective (the part warped anyway). Reducing HBP temp to ~50C (without chamber heating) was much more effective, causing the bottom layers to be less dramatically different from the upper layers, I guess? 

On Wed, Jan 16, 2013 at 8:16 PM, Michael Buffington <michael.b...@gmail.com> wrote:

therefore, the warping.

[Jetguy]

http://groups.google.com/group/makerbot/browse_thread/thread/8643712a29bc6a54/176c6586e2349bd3?lnk=gst&q=nylon#176c6586e2349bd3

> michael.buffing...@gmail.com> wrote:
> > therefore, the warping.

[Darrell jan]

Link seems to be broken.

[Jetguy]

There is a huge thread on printing with nylon with all the answers, I
provided a working link back to it and just checked it does work.

If it doesn't work for you (browser issue?), try using search in the
right uppper corner keyword nylon. I don't think we need 2 threads on
the same topic. All the answers are in that other thread and then
some.

On Jan 16, 11:44 pm, Darrell jan <darrell.l....@gmail.com> wrote:
> Link seems to be broken.
>
>
>
> On Wednesday, January 16, 2013 8:37:43 PM UTC-8, Jetguy wrote:
>

> >http://groups.google.com/group/makerbot/browse_thread/thread/8643712a...

[Michael Buffington]

Link is indeed broken (clicking it produces a 404, a server response), but I know the thread you're trying to point to.

It's fair to say it doesn't have all the answers - I've read the entire thread, added my input there as well.

It's also fair to say this isn't the same topic. That thread does indeed have useful information about 618 nylon and various stories of failure and success, but doesn't answer my original question.

That said - my design worked better than previous designs, though eventually it did warp. If the tendency of nylon is to want to shrink on its longest stretch, the strategy of reducing the thread distances is likely a good one. I suspect this strategy can also help with part strength as that's essentially what reduces the warping.

I'll try some variations on the design and share the results when I have them.

[John Driggers]

Michael - have you tried enclosed but heated to a more moderate temperature, with a slow cool-down?  I've had very inconsistent results, but - like you with the Nylatron I thought I had it licked a couple of times..

Jetguy - the other thread does not have the answers as far as I'm concerned, I've diligently followed it - and damned if I can get to it now!  Taulman has posted a lost of success stories, but I've yet to hear first hand from someone who's getting consistent results.

For the record, I probably get the most consistent - 1 in 5 without significant warping - printing on 3mm poplar heated to 50C, enclosed but not heated any more than what the bed and extruder produce, so runs around 40c inside.  .

[Clinton Hoines]

I find the same sort of. Enclosed with low heat 56 deg glass bed with hairspray and 236 deg extrusion. No fans etc blowing on the bed and the extruder fans diverted from the bed as I would ABS. Let the enclosure reach temp for about 15 min and cool down once done, very min warping on fairly solid larger parts.

 

Clinton

[Jetguy]

Not trying to be a smartass here but the title of the thread is "Why
does 618 Nylon warp?".

The answer is the same reason all 3D printed materials can warp. It is
entirely due to contraction of the printed material as it cools.
Because nylon has a higher rate of expansion than even other materials
we print with such as ABS, it may not be well suited for 3D printing
as printing as a material. Thus why most people do not use it and
haven't in the past.

And with that said, normal techniques that work with material with
lower rates of expansion might not be enough compensation to allow
nylon with a higher rate to effectively work the same.

What I can say is that the best answer is likely print slower. I mean
slow as dirt because most commercial machines print at say 12mm/s .
What that does is to force the filament in a local spot to cool and
shrink but not build up the internal stress over a longer distance,
thus likely to increase your chance of a print sticking and/or not
warping several layers up.

On Jan 16, 8:16 pm, Michael Buffington <michael.buffing...@gmail.com>
wrote:

> I thought I'd solved the issue of 618 nylon warping while printing by
> printing onto a Nylatron plate.
>
> Today, however, about halfway through a somewhat large print, it began
> warping right where I thought it might from having watched previous
> failures.
>
> It seems as if the warping is the greatest when long lines of thread cool.
> For example, the object I was printing had no straight lines for the first
> 10 layers aside from the minimal fill. The longest straight line was
> probably 1mm. At layer the first of 3 top layers began. Let the long
> straight lines, and therefore, the warping.
>
> Aside from creating objects that have minimal straight lines, is there any
> way of avoiding this? Would heating the chamber help?
>
> My current strategy is to try and see if I can trick the slicer into
> reducing line lengths.
>
> Here's an example - the first screenshot is the object's underside. This
> prints fine up until it's filled in on the top:
>

> <https://lh4.googleusercontent.com/-7wmACyqjoxg/UPdPhAsQT7I/AAAAAAAAAD...>

> Here's what I've done to try reduce the long lines - added 0.1mm slots:
>

> <https://lh5.googleusercontent.com/-sw-lojcuu80/UPdPz5Pi5CI/AAAAAAAAAD...>

> That gives me slices that look like this:
>

> <https://lh5.googleusercontent.com/-05LvBp_wGM4/UPdQKrezK1I/AAAAAAAAAD...>

[whpthomas]

This is almost the same question as "why do my TIG welds warp".

Translating what I know about welding to 3D printing, when welding I use a technique called stitching (not to be confused with cross stitching), this where I weld a short section left to right, as I stitch left

<-- Direction of travel

--4--> --3--> --2--> --1--> Weld action

So I am always working towards the tail of the previous weld as I weld. Thus the head effected area of the weld pool area is always approaching a previous weld that is cooling. It has the effect of physically bending the metal around the weld pool straight as it cools, reducing but not completely eliminating warping. Maybe someone needs to write module for skeinforge for Nylon and maybe even ABS that does a similar thing - it might be slower buy I suspect something like this would work. Certainly does for steel.

[John Driggers]

Jetguy - thanks for the valuable input.  As an FYI, at 12mm/s the heat soak from the heater-block leaves more residual head in the printed object than printing at 20mm/s does.  I think I've still got some FLIR footage we shot when trying to experiment with managing the expansion of 618.

I suspect that there is not going to be "one" right answer here, but there are a few of us interested in making it work - the material simply can do things that ABS and PLA can't.   Maybe the Makerbot list just isn't the right place to do it.

[Jetguy]

John, no I wasn't trying to imply in anyway to move this discussion
off. It seems self defeating though to separate answers in different
threads? This is just my opinion.
I'm not trying to be the thread nazi here. This is about trying to get
a better answer and I'm just steering the aproach. Sorry if that
offended somebody.

We have a specifc topic 618 Nylon printing. Specifically what is the
best way to do it. That's fine and even a good read for those not
working with nylon. Entirely possible a new technique or method or
part comes out of this that works for ABS too.

[Michael Buffington]

I've been able to print mostly warp free.

I went through a few design iterations that emphasized reducing the single straight lengths of extruded nylon, as well as added structure that would, in theory, help resist warpage. Here's an example from OpenSCAD:

I added the carved out slots thinking that they'd prevent Skeineforge from filling the area with long, highly shrinkable lines by forcing the extruder to fill in between the carved out sections with much smaller lines. Structurally, the voids don't hurt the part either - and possibly help increase its rigidity. I also sliced the part so that for every other layer there'd be a few more shells around the voids to increase their strength lengthwise. These perimeters would overlap previous filled areas beneath (which only had 1 perimeter), further "locking in" the void's walls.

Another design tweak is seen in the edge of the part - a fillet that surrounds the entire circular part. The fillet serve two purposes: it helps mate the part to another yet to be printed part, but also, in theory, helps reduce shrinkage. I have a feeling that the latter purpose might be wishful thinking.

I also sealed off the chamber with 1/4" plexiglas (way too thick, but I had scraps of it on hand).

Some print details:

• Speed: 20mm/s feed and flow rate
• 245C extruder temp
• Dimension: using the exact measured filament size of 1.68mm
• Dimension: Filament packing density of 1.0
• Dimension: 0mm retraction distance
• 2 deprime
  
• Inset: infill width over thickness of 1.45
• Inset: overlap width
• Fill: 2 extra shells on alternating solid layers, 3 on base, 2 on extra
• Fill: begin rotation of 45 degrees, 90 degrees on odd layer
• Fill: 10% rectangular grid infill (that percentage was hardly critical since the infill is mostly handled by extra shells)
• Comb: on
• Printed on a 1/8" sheet of Nylatron
• HBP heated to 30C mostly to help raise the chamber temp - it might make the Nylatron more flexible though, for better or worse.

Finally, the final print:

There was some very minimal warping that occurred starting at the layer directly above the end of the carved out voids. It's clear to me that they help. Even though I'd like to avoid warping entirely, the amount of warp for this print is acceptable.

Of note, the concentric spiral "teeth" on the top of the part ended up a bit messy because of the Comb plugin. When comb is activated, the extruder will try to avoid cross perimeters (I'm only about 80% confident saying that). As a result, the extruder went a big bonkers on those spirals. It spent a lot of time "trapped" in the maze they made, dropping little bits of ooze in random places. Without comb turned on I'm confident that these spirals would be much cleaner and 25% of the total print time would have been shaved off. They're functional, just not ideal.

[Adan Akerman]

Michael, that's a nice-looking part, way to go. 

I've started to wonder if hex fill may be less warp-prone than rectangular. By your logic, it helps reduce continuous lines across the part, right? But you do still have the perimeters.

I printed a nearly warp-free nylon part, and a pretty good-sized one at that, over the weekend. The big thing I think helped it succeed was that I drafted the sides of the base out at 45 degrees (image attached). Build surface was hairspray on glass. There are four magnets embedded in each of the parts. That worked very well.

Taulman's report of success with (unheated!) Garolite LE is pretty exciting: http://www.thingiverse.com/thing:43634. I've got a plate of that on order, should have it soon.

Adan

--
 
 

[hellphish]

Adan, what kind of hairspray are you using?

--
 
 

[Adan Akerman]

"Nuclear Hold," a.k.a. Garnier Fructis Style, Extreme Control, Anti-humidity hairspray, with natural bamboo extract, provides bounce-back hold, "EXTREME" (5 on a scale of 5).

I used to just spray a heavy coat (while outside the bot) and let it bake on while heating on the HBP. After reading suggestions here, though, I've started doing several thin coats before I use it the first time, letting the very last coat dry in the bot while the plates equilibrate. I've also started re-using plates without cleaning: just spray a new coat or two over the existing layers and go. I'll do that several times at least, until I have to soak a plate to get the print off. 

When dry, it's not sticky, just a hard clear coat. It does make a SQREEEEE when I slide my dial indicator over it, which is hard on the ears (and, presumably, not great on the indicator mechanism. But it's a cheapo, probably cost about the same as the bottle of hairspray, now I think of it).

I've been using HBP setpoint of 50C, though surely the glasstop is cooler. I've also got a generous coat of thermal paste between the heater board and the aluminum, which has significantly improved the system's ability to maintain a well-trammed state.

What about you?

--
 
 

[hellphish]

I print ABS and PLA on a layer of Aquanet super extra hold unscented. It works great for those plastics but is basically worthless for my 618.  Between a piece of birch, blue tape, glass, nylatron, and perf board, I'm having the best results with the perf board. The only downside is the little holes.

[Adan Akerman]

Did you see TBuser's comment about scratching the surface of his nylon print surface? http://www.flickr.com/photos/tbuser/8386869415/in/photostream I'm interested in trying that.

--
 
 

[Michael Buffington]

Great news!

I was able to print the part I was showing in previous posts with absolutely no warping.

I printed on a 1/4" sheet of Garolite LE (as seen here: http://www.thingiverse.com/thing:43634).

The first attempt printing to the sheet peel off the plate easily, but it's likely because I'd accidentally sliced it to print at 100mm/s.

The second attempt was printed at 60mm/s. I also flipped the part this time. I think the combination of the feed rate combined with the lower amount of surface area on the "stickier" plate was the trick. 

I'm looking forward to doing some more trials with it later tonight.

On Wednesday, January 16, 2013 5:16:31 PM UTC-8, Michael Buffington wrote:

[xenogea...@gmail.com]

On the McMaster-Carr website, the max temp of their Garolite products are listed as 235 F (~113 C) (except for the high temp G-11).  But it looks as if this 618 nylon is being printed at 245 C.  Is there any damage to the Garolite after a print from the heat?

[Michael Buffington]

No damage at all. In fact, it doesn't leave a mark.

[j.pickens]

I believe the temperature limit for garolite is mostly due to exposure to boiling water. If heated and wet, the water will intrude along the fiber matrix and cause delamination. Dry it can stand much higher temps.