Preform slicing speed


#9

The GCode for the print I mentioned n the original post was simply too large to even open up in a Word document, so I put a simpler model in Cura and slice’s it with the MOAI 0.1mm profile. it generated a 42MB Gcode file. The same model was then sliced in Slic3r for the MP Mini MSLA, and it gave me a 3.4MB CWS file.

FWIW, the MSLA CWS file is not G-Code it’s just a zip containing a print profile textfile, and a bunch of PNG files representing each layer. Here is what Layer 100 looks like

And here is an excerpt from the GCode with only layer 100 and the printer G-Code Preamble and post code. I had to convert them to a PDF because text files and such are not allowed here.
PM_nun_complete - Layer 100.pdf (137.1 KB)


#10

Yeah, the hardware should be fine

The slicing time is crazy though, even on my most complex prints I haven’t had it take more than 10 minutes


#11

You’re saying you never had a print upload take longer that 10 minutes?


#12

Same… I’ve had prints with >1 day of printing and it never took more than 15 minutes to calculate all the layers and then upload it to the machine… using my laptop.


#13

I’ll post a download link for the file later this evening so that you guys can try it and tell me what you see.


#14

As promised, here is a link to the file. The actual form file is 262MB, so I zipped it, and it makes for a smaller download (73MB)

Extract the FORM file, then open it up with Preform, select 0.025 and let me know how long it takes to upload to the printer.

Once printed you should have the 4 parts to assemble this:


...and now for something different
#15

What material profile did you select/ and can you please upload the STL files rather than just your FORM file

I presume from the picture that its your own supports rather than Preform’s - It looks like there is a lot of unsupported areas (the wings for example)?

Are you printing that exact combination of supports / orientation on the MP printer?


#16

The profile is Gray v3, and yes, the supports were generated in Slic3r, not Preform.

I exported the whole model with supports as a single STL file, then printed it both on the Form 1+, and the MP Mini at 0.025. without any additional supports.

All the areas that need supports have supports. The tips of the wings point upward at angles that don’t require supports. As you can see from the photo below, it prints just fine:


#17

Yes - The STL file (rather than the Form File) would have allowed people to play with the supports and see if its the lack of support generation by Preform that causes your problem (that is slow speeds in Preform upload)


#18

That’s the whole point though, isn’t i? If you let Preform generate its own supports, the output will be totally different.

The idea is not to judge its ability to generate supports, rather to see how fast or slow it is when it has to slice a very complex model, which in this case is complex because of the dense supports.


#19

Yeah, I’ve never had it take more than like 10 minutes, I haven’t done as large 25 micron prints, but I’ve filled the platform with prints at 50 microns


#20

So I take it no one tried to print this yet.

Oh well, it is a nice little model, especially for those that like miniatures. It can be printed with any profile (I think), since the supports are already built in, you can just drop it on the platform, select 0.025 layer height, whatever profile you want, and click print.

But if no one wants to try it, I guess we’ll never know how fast (or slow) Preform really is. Too bad.


#21

It’s not how large or how much area a print is taking or even how long it takes to print it. You can have a 5" cube, sliced at 0.025, and it will take a day to print, simply because there will be some 5000+ layers, yet the geometry is super simplistic, and the slicer should only take a minute or so to do the job and upload the model.

But then you give it a very complicated model with lots of discrete areas that need to be plotted, and you’ll see what the real performance is like.


#22

It seems you have printed it on Form 1+, well I don’t have a Form 1+ to replicate it and I don’t want to try printing your idea of support in the first instance. There are perhaps better ways that will print faster when used with the form 2. As you are so reluctant to publish just the STL files there seems not a lot of point for me to put any further effort into this…


#23

The files are readilly available from Thingiverse

But you obviously are unwilling to participate in the test I proposed. Whatever results you will end up with will not be representative of the proposed test.

Once again you missed the point of the test. It doesn’t matter what I printed it on, or with what profile. The idea was to give Preform a complex model to slice and see how long it takes it to do that one job, which would be measured by the amount of time it takes it to upload the job to the printer.

So enjoy the model but unless you’re willing to follow the given instruction you needn’t contribute any further.


#24

Is the test you propose designed to test how well and efficiently the Form 2 will do the print (with its own support generation), or to test how well Dudemeister designed his own supports and how quickly Preform computes with a file built directly on the build platform?

So enjoy the model but unless you’re willing to follow the given instruction you needn’t contribute any further.

Dudemeister - Does anyone need your consent or permission to contribute to a thread? Perhaps you might consider not being quite so arrogant to other people posting on the forum, none of us are superior beings and as such a little respect and consideration to other posters might go a long way.

You state that you are involved with IT, but seem to ignore a huge variable in testing the speed of an upload. That variable Is the one of if that upload is via Wifi, Ethernet or USB, all of which will give significantly different results


#25

True enough on the USB vs Ethernet vs WiFi issue. I hadn’t considered it since the Form 1+ only uses USB. So I suppose it’s via USB, which should be the fastest option (unless the Form 2 uses Gbit ethernet), although you’ll probably find that it won’t make much of a difference.

Withe respect to my comment to Dxxx, you’re right, it was condescending and beneath me so I appologize. It’s just that every time I posted something here and on other threads, he feels that he needs to take me to task and school me on the technology, terminology or just about anything that he feels he wants to argue or show his superior knowlede. Just take a look at the posts about GCode as an example.


#26

Well, I am sorry you feel that way Dudemeister. I apologise if you feel my opinion and posts offended your superior knowledge, that was not the intention. This Forum is a great place to share, learn and to see other peoples ideas or issues. But - I must admit, the first reason that I come on this forum is to look at Formlabs products, rather than to find evaluations of alternative printers.

On a personal basis I think the Formlabs products are quite incredible on many levels, sure, I see problems, but overall its helped bring 3D printing to the masses.

I agree with you that Preform is perhaps not as fast as some of the slicing software used by other printers, but still think Preform is a nice piece of software that has helped bring laser based SLA printing to the masses. It does what in reality is an incredibly complex mathematical task very well.

With regards my posts about G-Code, In the 1990’s, a large part of my day job was developing and working on interfaces for converting CAD drawings (both 2D and very early 3D) Often these drawings were generated by hand( lofting’s) on polyester film, they were then digitised in a lofting’s shed and then converted to G-Code in order to create a cutter path (that was controlled by G-code) I think that gave me a reasonable understanding of G-codes. Let’s put this into perspective in terms of technology, the 8086 processor only became available in 1978, so technology was developing at a fantastic rate in the 90’s (and continues to develop at an even faster rate today) I can still remember seeing and using Autocad when it had a dos based command set…

Today I am lucky, I get to develop 3D printers and materials as my day job, so, Dudemeister, you can see its a little more than a hobby for me. Again, I think that gives me a little understanding of the process and technology, but I also know that there are a great number of people out there with greater knowledge than mine.


#27

You know, if you would have stopped right after the third paragraph, I would have probably said that I agree with you that Preform is a great program (and I do believe it is), and left it at that.

But the rest of the post, is what I’m talking about. I don’t think I needed to know how you used to develop G-Code cutting paths from 2D and 3D drawings, and how much experience you have with 3D printers, in order to convince me and everyone reading these posts that you are right, and I’m clueless.

Oh, last but not least, “so technology was developing at a fantastic rate in the 90’s (and continues to develop at an even faster rate today”, that’s actually no longer true. Moore’s law is dead, and technological development in the silicon industry is slowing down, not getting faster.


#28

There are more important and entertaining things in life than debating with you. I don’t intend to be drawn into a debate of if, or when Moore’s Law will end, I will agree that has slowed down and the collision of Moore’s law and Rocks Law is inevitable. That is just one aspect of the progression of technology. In reality overall advances in technology does continue to develop and certainly is at an incredible pace, lets be fair - ten years ago 3D printing was the exception, now its common place and is taking industry forward at an ever increasing pace. Its opened possibilities that 20 years ago were only dreams.

So an example of something that is more important than a debate with you is shown below - I hope you enjoy it as much as I have!

image