Validating Isotropy: some thoughts on this excellent article

Read it here:

Here’s a diagram from the article:

All of this is my interpretation of what’s happening, from both the article and experience, so take it all with a grain of salt…

The tldr;/2 is: Before a model is fully cured, i.e. just out of the printer, it’s in what’s called the ‘green state’. Each layer’s shared surface still contains unreacted monomers. These later intermingle and react together during post-cure to fully harden the model into a chemically bonded solid. This green state is something FL has dialed in, and it’s why there are color profiles in Preform. Each resin gets to green, and behaves at green uniquely.

The diagram shows sharp edged brick-like layers with defined green ‘mortar’ areas, but I suspect it’s the most brick-like immediately at the PDMS-interface, then softening as the laser energy is absorbed and scattered into the layer. This rate of softening, or more precisely, non-hardening is likely a function of resin type and colorant to name only two obvious considerations.

So the green level of any resin is just a point on a multi-line graph, and each point makes tradeoffs. Keep it really green as a print, and then nail it in a UV curing over, and you’ll have a very solid part as all of the layers are now chemically merged to a greater degree. However, it’s probably more likely to fail during print, especially where tensile loading is highest.

But if you can minimize the amount of green during printing, it might produce a less isotropic part, but you may be able to take advantage of superior stiffness and strength of the object and support structure overall during the print - possibly in varying degrees between them - to achieve a desired outcome. Intuitively, I believe this will allow thinner wall thicknesses and finer detail, albeit at the expense of isotropy. Sometimes isotropy is less important.

This is why we as makers need knobs and gauges and levers and modular plugins in Preform. We need to, well frankly, help design the machine we’re using. Makers are typically tool makers in some form as well. There are so many things all of us will learn if we all had more freedom to explore and tweak and share our findings. A failure can be way more educational than a one-click-and-it’s-done kind of thing. I’m certain the software would improve at a much more rapid pace if advanced users had access to a mode that allowed more settings changes.

OpenFL is a great thing, and it’s awesome you’re doing it. Please also build it all into the gui - I want to experiment, but I don’t have time to do OpenFL right now. I think many feel that way.

1 Like

I’m glad you like the article. Be aware that for Form 1/1+, OpenFL includes a PreForm build with exposure knobs for your tweaking pleasure.

AH! Very cool. I had no idea about that. I will check that out immediately. Thanks @BenFrantzDale

Great. Let me know how it goes!

Hi Ben,
I had pulled the repo quite a while back, and looked through the files. I had assumed it was just Python stuff. I hadn’t noticed the links in the README to s3 for the modified Preform then. I’ve copied the two files mentioned. Is the exe to be run directly or is that an installer? What is the dmg file?

Also, I have recently been completely freaked by a failed firmware update - what is the worst-case scenario this firmware will do?

Thanks!

The dmg is a Mac installer; the .exe is, I believe a Windows installer. The firmware is backward compatible with regular PreForm. It just adds features for the Python tools to use, so for what you are doing, you should be able to just download the .exe (assuming you are on Windows), install it (it should install next to your current PreForm install, if I recall), and try printing with different settings.

One of the problems though with user testing is that the vast majority of users can only experiment with settings and just run tests to see what happens. Formlabs is better equipped to do testing, where they have tools to measure effectiveness and they can take the risk of something turning out bad where a user might ruin their printer.

1 Like

Exposure knobs?
Where are these located? Unless I messed up and am not running the correct version that I downloaded and installed, I can not find them. Is there a write-up on best practices on using them?
When was this new flavor released? Did you guys announce the release of this version on the other thread? I don’t remember seeing it.

True, and they can share photos and their experiences, and everyone will be richer for it. Formlabs only has so many employees, and only so much time to do testing. Why not leverage as many free worker bees as possible?

Maybe, but anecdotal data is data nonetheless. This anecdotal data can be duplicated by others, and suddenly it’s like real actual data :slight_smile:. Like “Hey, this setting produced very precise hole diameters for me in clear”, or “Don’t use this setting with Black, as it won’t stick to the platform, but for flexible it allows tiny support point sizes” - or whatever.

Personally, I prefer dimmer switches to standard switches, and variable speed drills to single speed drills. Know what I mean?

Hey,
I was going to share an image here that I screen-printed of my desktop, showing both preform 2.71 and preform 2.33 OpenFL running side by side, but apparently that’s too big, and I should post it to a cloud service sharing place and share the link. Really? A 1.9MB screenshot of my desktop is too big? To big for what? Who is the brilliant idiot who came up with those rules? Please go and fix that ASAP. That’s beyond ignorant. How about you go buy a bigger hard drive!

Seriously, if we can’t share images directly inline on the forum, WTF is the forum for? Unreal.

Oops, your post is too long. Please enter text at a cloud sharing space, and share links to the post here. Oops, link name is too long… Please go to tinyURL and reduce the link size to your cloud hosted text post and then share that… Oops, link is still too long, please create two posts with half a link each… Oops…

C’mon. This is BS.

Thanks for identifying some of the default forum parameters that are probably counter-productive. Glad we did away with thread auto-locking and I do suspect 2MB image size caps are a little overly restrictive. We’ve upped it to 7 so let me know if you’re have any trouble getting things uploaded.

Cool Frew. Thank You.

That’s pretty big for what most forums would support as an image attachment–a JPG even at a high resolution wouldn’t be that big

Huh? I think you might be misunderstanding the size of the image. It was less than Two Megabytes. A little more than a Floppy Disk’s worth of data. It’s a 19200x1200 jpeg image generated directly by win7 using the printiscreen key, pasted into MS Paint and saved. Standard run-of-the-mill screenshot. They’ve always been this size if your monitor resolution is this size and the quality is good. Nothing weird or bizarre here. Curious, what forums don’t allow a 1.9MB image?

I’m on a modeling forum that doesn’t even allow images because they think some people have bad internet.

In the OpenFL PreForm, you should see the “Load Custom Material…” button when Form 1/1+ is selected.

A reference file can be found here: OpenFL/Form_1+_FLGPCL02_100.ini at master · Formlabs/OpenFL · GitHub Start that and change things from there.

@BenFrantzDale The above does not equal “I’m glad you like the article. Be aware that for Form 1/1+, OpenFL includes a PreForm build with exposure knobs for your tweaking pleasure.” that you stated above. I am aware of how to load custom profiles. I was really hoping you guys made this a bit more user friendly with the “formlabs default profiles” as a starting point with sliders to increase or decrease exposure and tracking speed.

Sorry for the confusion; this is what we have at the moment.

I really liked the article, are you going to post the actual testing results in detail at all? This could be a great learning experience for my students.

1 Like

We just updated the article to include the raw tensile data. There should be a “Download the raw tensile data” section now. Here’s a link to the article or a direct download link for the data.

2 Likes