Polariscopy

[Ryan Carlyle]

It occurred to me today that a polariscope (see: photoelasticity) should work for observing stress distributions inside transparent 3d printed parts, just like it's often used to show residual stresses in injection-molded parts. This could potentially resolve a lot of arguments about extrusion strand cross-sections, warping stresses, and so on. If it works. 

I think a couple polarizing filters should allow a USB microscope to take pictures of the stress fields inside very thin parts printed in transparent plastic. I'm thinking single-wall or single-layer test prints coated in XTC-3D (or other refractive-index-matched coating) may work well. 

I have clear t-glase, esun PETG, polycarbonate, PLA, and nylon 645 on hand. I'm particularly optimistic about the t-glase+XTC-3D combo since the refractive indices match so well. 

[Ryan Carlyle]

Alright, circular-polarizing filters are in. I also have linear-polarizing filters on order. 

Esun test sample polycarbonate filament, containing some air bubbles:

Bent PC:

PLA:

PLA bent:

This is cool as hell. l can literally watch the color changing as I bend the filament. 

[Ryan Carlyle]

I know these pics don't look super amazing, I need to work on the lighting and rig up a filter mount to improve the color intensity a lot, but DAMN, I did not expect it to be this easy or this cool. I literally just stuck a piece of filament between two photography filters and pointed the microscope at it, and boom, stress-rainbows visible to the naked eye.

I need to try to take a video of the visible color changing in real time as I bent the print, that's really incredible to see.

[Brad Hopper]

Definitely fun stuff. Remember doing it in materials lab but now I have a reason and tools to do it. Just wish you could do the same with non-translucent plastic.

[Dan Newman]

On 14/05/2015 7:32 AM, Brad Hopper wrote:
> Definitely fun stuff. Remember doing it in materials lab but now I have a reason and tools to do it.
> Just wish you could do the same with non-translucent plastic.

Takes a bit more equipment....

http://www.sciencemag.org/content/344/6187/1013
http://pubs.acs.org/doi/abs/10.1021/jz502652n

Dan

[Petr Ptacek]

Yeah, somewhat harder to get it on Amazon then polarizing filters...

[Joseph Chiu]

Haha.  I used to work at RealD so I would just repurpose the circular polarizers off the glasses for experiments like that!

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/f3eaca01-a603-4d44-9e7a-2e89707d1863%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Ryan Carlyle]

Getting my polarizing microscope rig going. I'm printing an upper-filter-holder jig now. I'm hoping I can get a passing-decent video of color-changing once everything is immobile but the filament. 

Esun polycarbonate, bend and bubble:

Esun PLA vs polycarbonate from a "load filament" extrusion (after die swell from 0.5mm to 0.68mm):

Really interesting how manufactured filament, which is drawn down to proper dimension after extrusion, has a bunch of residual strain but a load-filament extrusion strand in free air has no observable strain at all. 

[Petr Ptacek]

Aren't they shock cooling fresh filament in the water? Few of the production lines I saw had always water pool right after the extrusion.

[Chris P]

The first water tank is typically warm/hot; the exact water temperature is an important component in producing filament with the correct diameter and circularity.

[Ryan Carlyle]

Super duper high tech stuff, here

[Dan Newman]

On 14/05/2015 8:11 PM, Chris P wrote:
> The first water tank is typically warm/hot; the exact water temperature is
> an important component in producing filament with the correct diameter and
> circularity.

Yup.

More importantly, it's the drawing down followed by non-gentle-cooling by the ambient
air (or possibly a final cooling bath) which bakes in stresses. The "filament load"
extruded filament wasn't subsequently drawn down (although it was subjected to the
ambient air which cooled it relatively fast).

In plastics manufacture, the drawing process is more often than not done to
increase tensile strength by orienting the long polymer chains post extrusion.
That it can introduce residual stresses usually not a concern. So, from the melt
a larger than diameter strand is extruded and then, for purposes of increasing
the tensile strength of the product, it is drawn down with the target being the
desired diameter.

Now, for glass fibre which is melt spun and drawn, controlling these residual
stresses (and inclusions) is very important since they can lead
to failures down the road. So, in the fibre optic business, this has all been
extensively studied. (When I was at the College of Engineering at Rutgers,
there was a large group studying this. They were even studying cooling from
cross air currents caused by the vibrations in the fibre as it was drawn.)

Dan

[Ryan Carlyle]

Ok, getting good stress-color images from single-layer PLA prints is actually pretty hard. I think the very small thickness of the prints is making the colors much fainter than they ought to be. 

First pass pics below. Esun "natural" PLA, printed at 212C onto a 30C bed, mostly 0.2mm layers by 0.4mm extrusion width. Mostly 6.4mm^3/sec. 

Random infill stresses around voids (purple), and mild shell stresses (yellow).

Stress where infill strands overlapped other infill strands (purple). This print was done at 200mm/s (nominal 16 mm^3/sec), note lots of under-extrusion.

Showing how shell stresses are axially aligned. Purple and yellow are equivalent stress levels, just different orientations. (I can swap the colors by rotating the filter.)

Comparison shot time! This is 0.2x0.4mm strands vs 0.2x0.6mm strands (same 6.4 mm^3/sec flow rate). I'm surprised by how similar they are. It looks to me like the wider extrusion width is higher stress, but it's very similar and I might be imagining it.

Rapid cooling with nozzle blower at 100% on 30C build plate, vs slow cooling with no blower on 55C build plate. The faster-cooled one shows higher strain. I haven't bothered figuring out whether the fan or HBP temp is more important. 

Note that these are single-layer test prints, so there should be zero warping stress and probably very low residual stress overall. The colors really show STRAIN and polymer alignment, not necessarily stress. I believe everything you're seeing here is residual polymer strain/alignment from the extrusion process that got locked in when the filament cooled. There doesn't appear to be any shear stress, just axial stretching along the strands. Needs a lot more experimenting though.

[Brad Hopper]

Hey Ryan, I assume these single layer prints are still adhered to the glass on which you printed them, not lifted and then placed side by side?

[Ryan Carlyle]

Lifted off a PEI bed and then placed side by side on the microscope. I let them cool all the way down before removing from the bed and gently spatula'd to make sure I didn't add any new strains. 

[Ryan Carlyle]

I tried some single-wall box vertical tests, but they're surprisingly uninteresting. I was expecting to see some kind of nifty shear pattern due to warping stress, but no, everything looks perfectly uniform once the print gets more than 5-10mm above the build plate. I'm wondering if linear filters would show something different from circular filters. 

[Ryan Carlyle]

It occurred to me that the strain/drawing crystallinity induced in the plastic during extrusion is probably the biggest factor in optical transparency. Amorphous plastics can be transparent, crystalline plastics tend to be opaque. Stretching out the polymer makes it more crystalline and less amorphous. So when we print PET or PLA or PC fairly fast and get cloudy extrusion strands, that likely means the high extrusion strain is inducing a lot of crystallinity in the material. Which means the transparency/opacity of the extruded strands (in thin part sections) can be used as a quick visual check of how much residual strain is locked in the part. That, in turn, is a predictor of stiffness/brittleness. Optically-clear parts should also have high toughness and more flexibility than cloudy parts. This fits with the annealing experiments I did a while back -- after annealing, PC gets extremely transmarent and much tougher.

Of course, none of that works for opaque filaments. Or with thick parts -- the corner voids between strands also cause light scattering and cloudiness. But I think this could be a decent rule of thumb for printing test coupons to dial in desired properties with materials like nylon that exhibit significant strain-crystallization strengthening. 

[Ryan Carlyle]

Hadn't occurred to me to try this until last night, but drive gear tooth marks look really cool.

[Ryan Carlyle]

Roll-screwder r3 bite marks (on polycarbonate):

[Ryan Carlyle]

Ohhhhh, this is just lovely. I'm catching "cap zone" behavior during a filament load sequence. Plus some fascinating drive gear bite behavior.

SCIENCE!

[Bob Bilbrey]

Congratulations .  An award winner, Ryan!  

What USB microscope + polar filters are you using?

[Ryan Carlyle]

I posted a photo of my setup a little while ago: https://groups.google.com/d/msg/3dp-ideas/C_DQzpgr8Pw/g1s9SjNCdMAJ

The microscope came from Adafruit, I don't know what brand it is. https://www.adafruit.com/products/636  You can get cheaper ones or better ones on Amazon but the Adafruit one is pretty good.

Microscope stand: http://www.amazon.com/gp/product/B004M8SQ7A

Circular polarizer: http://www.amazon.com/gp/product/B00004ZCC1

Note that you need two circular polarizers, I took the element out of one and put it inside the microscope stand on the light source. The part you want to look at has to go between the two filters. 

[Dan Newman]

On 28/06/2015 9:01 AM, Ryan Carlyle wrote:
> Ohhhhh, this is just lovely. I'm catching "cap zone" behavior during a
> filament load sequence. Plus some fascinating drive gear bite behavior.

That's quite nice. PCA as well?

Dan

[Dan Newman]

Typo:

... PC as well?

[Ryan Carlyle]

PLA. This spool of Esun natural PLA has a really great residual manufacturing stress that helps indicate the heat-affected zone and highlight the tooth bite marks. But I should try some t-glase and PC too... 

[Brad Hopper]

Really cool. Thanks for sharing that.

[Ryan Carlyle]

I need to repeat this for a PTFE-lined hot end, ulgh. What's the minimum effort I can expend to rig up a "parts bin" FlashForge mk10... 

[Dan Newman]

On 28/06/2015 1:26 PM, Ryan Carlyle wrote:
> I need to repeat this for a PTFE-lined hot end, ulgh. What's the minimum
> effort I can expend to rig up a "parts bin" FlashForge mk10...

Min effort tends to lead to "time == $$" therefore minimize time. IOW, buy
a Mk9+ unit from, e.g., makergeeks ($90 + shipping), wait for it to arrive,
disconnect your own wiring from the motherboard, connect the makergeeks unit, then do some
extrusion with it. Don't physically connect it to a bot; do check the
cartridge resistance first.

And I don't know for a fact that that Mk9+ unit has a PTFE liner....

Dan

[Dan Newman]

Also, some of the direct drive units on eBay list a PTFE liner. They also
have thermistors and 12V fans & cartridges. Real question is what pinch
gear. Plenty of the direct drives on eBay seem to use a much smaller diameter
pinch gear -- ~7mm instead of ~10 - 11mm.

Dan

[Ryan Carlyle]

I have a Monoprice mk10 bar+hotends pre-terminated to fit my R2x swappable extruder assembly. Relic of experiments long-past. Just had to rig up an extruder, heatsink, and fan. 

Naturally, the hot end I picked jammed within seconds of extruding. %$#@&

[Ryan Carlyle]

Alright, PTFE-lined hot end is not significantly different. Somewhat longer transition zone.

 

[Ryan Carlyle]

The 2mm diameter cap zone has trouble fitting through the inlet to this thermal barrier tube. You can see a little bit of shaving in the pic above. Oh well. 

[Ryan Carlyle]

This is t-glase:

I had trouble getting the t-glase and polycarbonate cap zones to pull out of the hot zone without damage, most of the test samples ended up looking like this where the high-shear semi-molten region got partially pulled off the unmelted solid core by the thermal barrier diameter step: (more t-glase)

PC: 

Unrelated to cap zones, I got this really nice shot of polycarbonate starting to progressively strip when I fed it into a hot end that was too cold for the material. (240C I think, I usually print that PC at 265C.)

[Joseph Chiu]

Go watch a 3D movie at a RealD 3D theatre and pocket the glasses on your way home.  They are pretty good quality circular polarizers. (I used to work for them.)

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/d6a0df27-077c-45eb-9a34-15cf3b1b8071%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ryan Carlyle]

On Monday, June 29, 2015 at 6:33:54 PM UTC-5, Joseph Chiu wrote:

Go watch a 3D movie at a RealD 3D theatre and pocket the glasses on your way home.  They are pretty good quality circular polarizers. (I used to work for them.)

Hmm, I didn't know they used circular polarizers. Any idea why they do that instead of linear polarizers?

[Joseph Chiu]

It's so that the extinction ratio is high even when you tilt your head.  As
my manager at the time used to say: it's so your wife can rest her head on
your shoulder and still enjoy the show.

There is a tradeoff to going CP -- it screws with the color a little bit,
since the retardation angle is a function of the wavelength.

Anyway, now you know where you can grab more CP's easily for science! :)

Oh, and of course, if you need to do some quick linear tests, just use the backside of the CP's.

On Mon, Jun 29, 2015 at 7:18 PM, Joseph Chiu <joe...@joechiu.com> wrote:

It's so that the extinction ratio is high even when you tilt your head.  As my manager at the time used to say: it's so your wife can rest her head on your shoulder and still enjoy the show.

There is a tradeoff to going CP -- it screws with the color a little bit, since the retardation angle is a function of the wavelength.  

Anyway, now you know where you can grab more CP's easily for science! :)

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/1838812e-925b-40ec-b496-7ee95c573a6a%40googlegroups.com.

[Ryan Carlyle]

Makes sense.

What I should really do is set up a monochromatic light jig. That would let me take photos that would show stress bands with enough contrast for black and white publishing in my book. (I'm hesitant to commit to content that requires color dead-tree book publishing, since high page count color books are ~3x more expensive relative to black and white books.)

[Ryan Carlyle]

I forget who wanted to see it, but here's the PLA behavior with a really sharp drive gear (Tatsu v2):

Seems about the same as the others. Bite mark still seems to disappear when the plastic hits the annealing / glass transition point. 

[Ryan Carlyle]

Another theory is that the force required to pull the filament back out of the melt pool actually "straightens" the polymer molecules, drawing the filament out a bit so the bite mark gets smoothed away. Dunno. 

[Steve Johnstone]

Amazing images Ryan!

[Ryan Carlyle]

Science is magic :-)

[Ryan Carlyle]

I really hate PLA, because it breaks in my filament feed path after a few days of not printing. Brittle materials that creep at room temp are bullshit. Probably doesn't help that there's a ton of residual extrusion stress locked into my Esun PLA, as you can clearly see from the bright colors on all my polariscope shots of PLA filament in this thread. Not sure if other brands are better about that or not... would be an interesting comparison point, but I don't feel like spending a bunch of money to buy a bunch of filament I don't actually want to use for printing. 

Anyway, I took some shots comparing the end of filament broken overnight just sitting in a feed tube, versus other methods. When two pics are shown, they are opposite sides of the same "end" from the cut. 

Remember, each specific color is one stress orientation and magnitude. 

Feed tube creep failure has some interesting crazing / strain bands:

Scissors:

Small flush-cutters:

Large diagonal cutters:

Razorblade:

Pulled apart in tension by brute man-force:

I like the tensile failure one the best. 

[Ryan Carlyle]

Ran into some >2mm bulges in my Esun PLA filament. They would've wrecked two 12 hour prints if it weren't for my filament monitor. 

Bulge 1:

Bulge 2 had some zany contamination:

Bulge 3 was baaaaarely big enough to get stuck:

This is pretty unusual for Esun. 

[Joseph Chiu (Toybuilder)]

Unusual, but does happen... Glad the Tunell monitor is working as intended!

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/5e652686-73c9-4a97-9b4b-010976aaa030%40googlegroups.com.

[Ryan Carlyle]

Love my filament monitors :-)

I think I'm going to have to stop ordering PLA made by Esun though, it's driving me crazy how it breaks in six places in the feed path if I leave the printer alone for a few days. As more colors of PETG become available, I get less and less interest in printing PLA. 

On Wednesday, January 13, 2016 at 8:25:38 PM UTC-6, Joseph Chiu (Toybuilder) wrote:

Unusual, but does happen... Glad the Tunell monitor is working as intended!

On Jan 13, 2016 5:38 PM, "Ryan Carlyle" <temp...@gmail.com> wrote:

Ran into some >2mm bulges in my Esun PLA filament. They would've wrecked two 12 hour prints if it weren't for my filament monitor. 

Bulge 1:

Bulge 2 had some zany contamination:

Bulge 3 was baaaaarely big enough to get stuck:

This is pretty unusual for Esun. 

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[whosawhatsis]

Yeah, I don't know if it's just that I've been hanging around more printers that spend more of their time idle, but I've been hearing increased reports of filament breaking like that recently as well...

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/b04a4f57-c3ca-45c4-b0c2-b71b62aeba40%40googlegroups.com.