3D STEP File Compatible: For supports export and import file

REASONS:

1 - This is for support modifications / improvement originated from Preform and edit it back in Solidworks, Creo (solid CAD) or 3D Max, Maya & etc. (surface CAD).

As the Preform supports generation are too limiting, meaning the Preform should be allowed to export as STEP, STL or OBJ along with the supports for customisation.

2 - In Preform, there will be many benefit of using 3D STEP beside using the outdated STL files (so last century):

a - Much smaller file size
b - True geometric shapes & details, which unlike STL that uses polygons to construct an approximate 3D mesh structure. No other industries uses STL anymore like CNC or CMM or CAD CAM or PCB gerber file or Laser cutting (vector based).
c - Quicker to load
d - Much easier to repair or modify. As STEP file is generated from a true solid CAD from engineering CAD software like Creo, Solidworks, Unigraphics. STL is a burden not a benefit for engineers.
e - Benefit to Formlabs as well, as they can have access to open source library like CrabCAD and other free 3D CAD library for downloads IN WHICH MAKERBOT IS ALREADY DOING.

3 Likes

Thanks for the suggestion! Iā€™ve forwarded this to our Software Team so that they can look into it.

My understanding is that an STL or similar file format like an OBJ is required by printing software to generate the correct toolpaths. The STEP format provides the requisite data to fully define a 3D object but most 3D printing programs will require a point based format to convert the object into the actual toolpaths.

It could be possible to import STEP files directly into PreForm and then have them converted to an STL esque format and sliced under the hood. This could be done non-destructively so that a STEP file could be exported after support generation though I believe the support structures and rafts are mesh objects that might not be especially easy to work with.

1 Like

STL isnā€™t required by printing software, however most printing software uses STL. Thatā€™s because itā€™s a non-proprietary standard that is very easy to support in software, using math that most of us learned in High School. And almost everything can read/write STL.

STEP (and to a slightly lesser extent IGES) are much more complicated to support and much less universally supported. They generate smaller files because the geometry is more ā€œabstractedā€ (a curve isnā€™t approximated by multiple discrete points like in a STL, itā€™s got a start point, an end point and a radius). But ultimately, as Frew says, a STEP would need to be converted by PreForm in to something very much like a STL since thatā€™s the native format the printer is designed to interpret. And Iā€™d argue that for what the printer does, a different native format would not result in improved prints. STL is more than sufficient to meet/exceed the resolution of the Form2.

Iā€™ve never had a situation where I needed to edit supports more accurately than PreForm lets me place them. Iā€™d be interested in understanding an example use case if you have the time to explainā€¦

I do have situations where I find a model file on a site like GrabCAD and itā€™s not available in STL. But for those situations, I simply import it in to one of my many CAD and/or format conversion programs and export the STL I need for the printer.

People only use STL as a last resort, not because it easy to use (it not).

During the late 90s Engineers have opt out STL and in favour of STEP (less so for IGES) as universal standards.

The only reason why STL still use today because Artist (unlike Engineers) couldnā€™t agree on a universal file standard until few years ago of using obj, This is because in the past, content creators (Artist & Engineers) rarely export from one CAD users to another as it was all done under the same roof in those days.

Artist use Obj and Engineers use STEP and it unreasonable to think Formlabs only aim for artist and not engineers who also heavily involved in content creation.
Because STL will always be an approximation not true geometry, in fact, CNC, 2D laser cutter, assembly robots, NC programs ALL uses STEP (for 3D) or DXF (2D) to be convert from digital information to physical manufacturable parts

This is because STL is highly inefficient i.e. highly compressible (too many repetitive structure), even if crank the settings to really high details (high polygon counts) you will end up having a very large file size. Most of my STL files end up 50mb~ 500mb eventhough Formlabs informed me there limit is around 200mb before the software crash.

In my book this is unacceptable, because Iā€™ve on some rare occasions I had to compromise between built in design features geometries OR retain those features and reduce the resolution to the point where I can physically see individual polygons on the prints.

It is true STEP is more abstract, but any standard files like STEP, parasolid, Iges, STL obj and etc need to be converted into toolpath anyway (g-code / .form) this includes STL. What Frew saying is that Preform is currently only written to read out STL by default. If you use any professional software like Solidworks or 3D max or Maya you will notice any file require converting to display meaningful data on screen (other than STL), but the time it takes to convert is no differences to the time it takes (or even faster than) to read out all the XYZ positions from an embedded STL.

Iā€™m not sure if you know this, STL actually have 2 variant ASCII and Binary. Binary is the slightly more efficient and slightly smaller, but if you save a simple cube in CAD into ASCII STL and open this file as note you will physically see all the ZXY position necessary to create a cube.

The conversion program is nothing new nor complicated, in fact there open source out there to convert STEP into G-codes (actual machine codes in which the machine can understand).
http://replicat.org/generators.

Also, as I understand it (I could be wrong), Preform was written based on Autodesk (they specialised in vector based file in their DXF), which is completely different to why STL works. This is why Formlabs suggested Meshmixer for STL modification.

Preform supports generation are good enough for most situations, but are too limiting if compared to professional SLA support customisation. Also, meshmixer seem very un-intuitive if compared to any Engineers CAD e.g. Soldiworks, Creo/ Proe, unigraphics, Catia and etc.

Note: Iā€™ve starting to noticed for the past few years a number of professional 3D artists (content creator) uses 3d max / Maya for freeform modelling, but also uses solidworks alongside (or other engineering CAD software) to create geometric model as they can use it to measure (true) and modify geometry using STEP before importing into 3d Max.

Even if it a polygon/ mesh based supports back into STEP, I could create a whole configurable library with it in CAD and manipulate the geometry using this library suing configurations. Because there are some situation Iā€™ve come across where supports were generated on an inaccessible / hard to reach location in which I definitely donā€™t want support attached to certain surface. Creating a landing pads on the model is a quick fix, but not a true representation of the 3D model.

Preform should not only supplement obj or STL, but should seriously consider STEP compatible, because most 3d content creators uses STEP (and some obj).

This is the reason why most 3d engineering model library from ā€œGrabCADā€ or ā€œ3D content centralā€ uses STEP, parasolid & IGES, but rarely STL, as it not a true geometry or easy to measure. As Iā€™m both CAD creators (artist & engineers, used both for many decades) I know everything being produced in CAD are actually all true geometry before downgrading / converting into approximate STL.

[quote=ā€œFrew, post:2, topic:14332ā€][/quote]

As I understand it, the point based system is actually based on G-codes to create a specific toolpath and not necessary have to work with STL as it is a triangle / polygon point based. This is seen in all CAD CAM like NC program, even in 5 axis CNC. In fact, if you create a cube from CAD it create 8 corners or points in 3d space, but how STL works is done by triangulation i.e. placing a diagonal lines on all facets, which is why NC toolpath are not normally done in STL.

www.youtube.com/watch?v=btRd9Sb9CAM It showing a gantry system, but it can be translated into Delta or Robotics arm like ABB or galvo mirror movement in which Formlabs is using. The most important aspect is to allow the machines to understand the physical geometry in 3d space, moving from point to point (toolpath) is something that is generated in the software, but the points used in STL has already been modified from the original CAD model. i.e. STL file doesnā€™t contain toolpath information.

To keep it short, and beside using STL and OBJ as imports, I strongly believes the entire engineering communities would appropriate this compatibility feature with STEP Import & export as we see STL as a dud file or read only file, As I cannot easily modify or measure accurately with STL, but instead creating a whole model from scratch would be easier (for geometric models, not freeform models).

It true meshmixer has this ability to modify STL, but it is not intended for creating geometric shapes like the complex parts within an assembly (artist CAD donā€™t normally have assembly feature) which normally uses engineering tolerances of +/- 0.1. This is to ensure everything fits after prints.

We need an alternative file type (beside STL) that could truly match your current machine capabilities i.e. print tolerances.

Regards

K

It makes sense that the STEP format would allow you to represent greater detail more efficiently as compared to STL or OBJ formats. The current slicers (at least in the 3D printing space) seem optimized to convert the triangulation based formats into toolpaths but it should be possible to do something similar with STEP. An easier first step would be to allow for STEP import and then convert to triangle based format under the hood so that the current slicer doesnā€™t need to be modified. A more thorough development might be to convert the STEP format into toolpaths directly although I donā€™t know if there are edge cases that make this challenging. Thanks for the suggestions! Our team is actively looking into this.

STEP and IGES exist as a common interchangeable file format, designed to allow different CAD programs to exchange editable models. Most CAD programs such as Solidworks, Autodesk, Rhino, etc, all have proprietary file formats, and in general, one program canā€™t read the otherā€™s native format. Thatā€™s where the ability to export to a common format such as IGES or STEP comes into place.

The reason most printing services ask for the STEp file is because they sometimes need to edit or make small changes to the model, either to fix small problems or make the file printable or in the case of injection molding, to add or modify necessary features like drafts.

Additionally, some CNC, or injection molding machines support input of STEP files directly. The more sophisticated ones will convert the files to curve driven tool paths, but a large majority of CNC machines still use coordinate based G-Code, which means the machine first converts the model to a tessellated shell (STL), then generates the G-Code.

STL files are generally not very editable, at least not the in the original context, as the individual parts, objects, primitives and features are lost in the export process. So one reason why you would want a STEP file, is to import it into a CAD program for further editing, but for printing purposes, itā€™s not required, nor desirable in a lot of cases.

I donā€™t know of any programs that can import a complex STL file and convert into something that is editable as the original was. Programs like MeshMixer are more like clay sculpting, than solid modelling CAD.

BTW, a 200MB STL file, will have about 4 million triangles. Those triangles would be way too small for anyone to see the individual facets.

Nurb files are resolution independent unlike poly objects similarly to postscript files vrs bitmaps.
Cooking the files down to obj (supports for multiple meshes in a single file that allows overlaps) and stl files to size will not effect the resolution of the print since the printers canā€™t print the details that you can see from a properly exported file. Much like printing a high res bitmap to your printer.

I think there is also a licensing issue with the nurb formats which would make the software cost prohibitive. Interpretations (how well step, iges, 3dm) are implemented into the software can cause geometry errors such as co-planer faces that can cause iterations in paralel surfaces causing striping.

Next issue is supporting the visual feedback and video card requirements to see the objects on screen. Even though you may be working with solid geometry the gpu is actually rendering out polygons. Using nurb files would require higher end systems to handle them where the same file in a high resolution polygon obj or stl would have a higher frame rate to be able to manipulate.

Lastly I had asked for the ability to export out files from pre-form as stl or obj without the supports. I would think it would be unfair for Formlabs to produce software that is free only to have people use it for competing products. Iā€™m sure they pay a pretty penny for the source code and need to get revenue back from their users.

If your having bad exports from Solidworks you may want to snag a copy of MOI3d that has nice controls on tesselation for export with visual feedback. I use both MOI and recently FormZ pro and though FZ does beautiful exports, MOI has a slicker export.

Iā€™m not knocking Solidworks but itā€™s intention is focused on the modeling side and not so much on export.

If I recall Nurb file was originally developed by 3D Max in the Artist CAD industry, but later adopted into an universal standard called IGES for the Engineering CAD industry.. Whereas Artistic CAD industry couldnā€™t agree on a universal standard between Maya and other smaller bespoke artistic software (used in the film industry) in the late 90s so they starting to adopt IGES in the early 2000 before a takeover by Autodesk who already own 3D max.

Same can be said with Engineering CAD such as Solidworks and ProE (now creo) is actually written / originated one of the same, before branching out into 2 separate companies. Many base features remain this including file structure.

Since early 2000 many Pro 3D CAD users from both sectors (Artist & Engineer) saw the divergence of Pro 3D CAD packages and there was a growing demand for a universal file standard for cross party communications. i.eā€¦ When a Pro designer uses one package e.g. Solidworks / Unigraphics and require to be manufactured (sub-contracted) by another company for injection moulding for example they most likely uses other CAD packages e.g. Catia / ProE (superior at mould tool manipulation),

In short, the universal communication standard (for Engineering secto) has already happen / setup since the early 2000 using STEP (less so with IGES), because before 2000 most still uses AutoCAD (2D) or worst pen and paper.

What IGES specialises at (as part of Nurb structure) is to retain freeform surface geometry & 3d sketch & splines from the original CAD you have created. This means 2 things: It allow for open ended or missing surfaces during translation whereas STEP is a closed loop, because STEP is a true solid by default whereas IGES (or nurb) isnā€™t as it only retain surface information and not interior. So what happen with Nurb type file is that when you import file like IGES back into CAD is that they try to stitch all the exterior information back up to make it into a solid cad object again.

This can only be done in Engineer CAD and not Artist CAD with Nurb files. it two very different thing, in fact it 2 different CAD category, because by default ALL Artistic CAD software are written as surface only (no solid information inside as there is no need for creative or even 3d printing industry), whereas Engineering CAD are more complex where by default it create solid information (not hollow inside) but it can also import surface only information as well. This is why Engineer CAD have many ways to measure & dimension an object e.g. centre of mass, density of object (useful for 3d print with supports) parabolic arc generated by specified mathematical curvature e.g. aerofoil design. Whereas Creative / artistic CAD will always be created by approximation, which is fine because I cannot see any reason why Creative CAD should adopt solid creation considering it hold more unnecessary on screen information compared to STEP, parasolid & other Engineer CAD file.

But this so called unnecessary on screen information, is actually very useful for turning digital information into physical object (whether if it injection moulder, sheet metal maker, investment caster, CNC machinist & etc ), this is why manufacturer uses STEP as a standard (for 3D) for imports and not OBJ or STL. In fact, speak to any Professional Model makers (who may own upto 20 different 3d priners from Pro to Enthusiast machines) and majority would recommends STEP not OBJ or STL, because it allow them to easily manipulate a true solid object in CAD compared to ā€œsurface onlyā€ (stitched) before printing.

I know all this because Iā€™ve been using 3D CAD and 3D printers professionally & continually since early 2000s and Iā€™ve used both CAD discipline (Creative & Engineer, 2 each) this including numerous 3D printing machines and research into possible applications into the real world for various companies .e.g Rapid Tooling, Jig & Fixture, traditional model making for clients & etc that just within the 3D printing worlds (a.k.a. additive manufacturing).

My main responsibility is to ensure the digital design is good enough to be mass produced and I failed to see any benefit of using CAD packages . from the artistic sector such as MOI, Blender & etc. especially when this MOI is a strip down version from Pro software like 3D max / Maya. I can see you are quite familiar with Nurb type 144 and even 126 in which IGES is excel at of extracting vector / points information, but STEP file by nature is more stable and less error.

The point is every time you convert a file (i.e. importing or exporting) from native from your CAD, original details are lost, this is especially true with STL as it is converted into vector based (just open a STL ASCII based file in note and you will know what I mean).

It like saving RAW (native CAD) into JPG (STL), whereas lossless file (almost) conversion like RAW (native) > Compressed RAW with Zip / RAR (STEP, Parasolid, IGES)

For Engineer CAD, what you see on screen isnā€™t what you get as you can get much more detail, which unlike Artist CAD. In fact, I would argue a 20mb native model (converted from STEP) is far better for most PC to handle this workload than a 200mb STL. and has very little with the GPU spec.

It true that since 2006~2012 Iā€™ve noticed a further divergence of open source 3D packages derived from Art CAD sector (vector / surface based) and less so from engineering CAD, because it easier to code surrounding STL.

It true that STEP isnā€™t as straight forward as STL when writing the software, but STEP is actually open source and definitely not propriety properties to any company.

http://opensourceecology.org/wiki/List_of_CAD_Programs

I didnā€™t buy the Form2 machine just creating pretty looking models (this isnā€™t the only applications), but retain some of the function for assembly for example. Because of this 200mb limit, I on occasion had to compromise with the decision of whether to reduce the polygon counts to visible sees rawness of the prints OR reduce function of the models, gears threads, piston, o-ring, body within body (not with Form2). This is to ensure everything fits and sometimes assembled with bespoke metallic components, without too much hassle of post-modding the prints or the CAD models as we donā€™t have all day for that if compared to the Enthusiast communities. STEP would alleviate that when time is of the essence.

Beside using OBJ & STL, Even Makerbot (another Enthusiast based machines) already adopted STEP so why donā€™t Fromlabs? As it would open up CAD users from the Engineering side as well, such as supports manipulations (beside the Creative / Artistic CAD users). Or even using CAD library like GrabCAD and 3D Concept Central (most models in here donā€™t use STL for the reasons stated above).

Regards

Yes completely agree, especially about CNC = Gcode (or even PCB = Gerber), but once the native file is converted to STL it no longer retain the likeness from the original.

Anyway, I sometimes have to crank up the polygon count to 200mb+ if it a very complex object to print (for fit and function not just for aesthetics) especially with seals for enclosed vessel testing and sometimes I have sacrifice aesthetics. So I end creating 2 different models, for function e.g. testing or aesthetics. So 4 million triangles is nothing when I have multi-facet bodies that merged into one this is especially true with multiple functioning threads (knife edge scenario).

Iā€™ve used a few NC program in the past so to optimise between toolpath and design of the workpiece. I generally liked the way how the Gcode manage (somehow) in handling mesh polygon as true circle for cylinder for example (somehow? curved toolpath?). Also, just spoken with an external supplier of mine who deal with most of our modelling making besides Pro 3D printers they also have CNC and they say they still prefer using STEP (or native) over STL as it gives them the flexibility to modify when fit and tolerances are needed for re-adjustment.

Iā€™ve also used CMM as well and Iā€™m certain they donā€™t use G-codes. Also used 6 axis robotic and I know they still have option to use the ancient STL ASCII, but with 200+ points for the robot go to through in an assembly line is insane, where these production PCs couldnā€™t even handā€™e this kind of load.

In the Artist / Creative CAD, it is quite common to edit STL, but less so in Engineer CAD, but you can at some extend edit it in Pro-e and I believe Catia as well, as both of these have a much more extensive tool kit for surface / vector manipulations in a controlled manner (Artistic CAD donā€™t, e.g. 3D max, Maya and all the free ones).

1 Like

Even if nurbs is the base type of obects used (non rational b splines) but in addition to the geometry type there are other information stored in files such as step, iges and 3dm that are program specific. Some of that information isnā€™t necessary such as parametric history etc.

I think the bigger issue is what would be the minimum requirements for PreForm? As it stands not it is very feasible to be able to print models via a tablet or ipad/iphone if they chose to port the code over. Not likely to be done if they have to start deriving layers and interpreting the slices from solid geometry.

I use clipping planes quite a bit in FormZ Pro and love them but on my laptop the clipping planes can grind the poor thing to a hault as it is crunching away to display the new visiable geometry.

Having a file cooked down to itā€™s lowest form ie obj or stl makes it much easier to handle and drastically lowers the computations required to display the object as well as calculate stress points etc. As I said before if your losing detail on your exports then either your exporting at the wrong tessellation settings or the program does not offer the options for tessellation or lacks the controls to do so.

Make sense?

SolidWorks actually handles a pretty big list of formats, including Rhino. I donā€™t have a copy of ProE anymore, but my recollection is that it also had a pretty decent list of importable formats. Usually if SW wonā€™t handle a particular format, I can find (or I already have) a program that will convert the file in to something SW can handle. Or vice-a-versa.

I understand that, and I know Soldworks can imports a bunch of different formats, but it can only do so with limited success, and only if the program it imports from shares the type of modelling architecture (solid modelling). Try importing a model from a surface modelling (NURBS) program, and youā€™ll find there are no readily available tools.

Hereā€™s a good example 3D Studio max to Solidworks (or vice versa). I donā€™t know of any conversion programs.

Moreover, even within solid models, export and import do not always yield good results, or rather properly editable results.

STEP format in particular is not very good at preserving the original workflow.

Take a more complex Solidworks model, then export it to STEP. Then re-import it. it will open the STEP file, and then it will attempt to do a feature recognition, and 85% of the time, the model will not be made in the same way it as originally made, meaning the work flow used to arrive to the final shape of the model will be will be completely different. Not only that, but some of the more complex parts will not be recognized, it only be ā€œimportedā€ without a means to modify them.

2 programs come to mind if your converting or trying to edit various models. Rhino is a good choice and has nice repair tools and ability to convert point clouds into surfaces etc. I had used Rhino and fount it a little tedious but it is a rock solid product. I recently jumped much of my work over to FormZ Pro which is a hybrid modeler that has sub-d surface tools in conjunction with nurb conversions. I found FZ to be pretty powerful since it does have some parametric abilities. Nothing on the scale of Xenon or Solidworks etc but enough to give you flexibility to design and modify products for 3d print. One of my favorite features in FZ is the clipping plane that allows me to slice objects and still work on them to see fit before I print. So far the exports from FZ have been coming into PreForm error free (cross my fingers and toes). I did find Rhino to be crash proof, FZ i recommend saving often.

Please look at the link provided by the 2 CAD sectors.

Art CAD: also lookup in this site STEP

Engineer CAD:

STEP (including all other Native Engineer CAD format) contain original information, whereas STL donā€™t.

Because when drawn / create 3d stuff in ā€œArt CADā€ it only contain surface information, whereas ā€œEngineer CADā€ by default it create solid 3d objects. So when using Engineer CAD to save into native, it retain that solid information and when export to IGES or STEP it tries to retain all that solid information when compressing, but if export from any native to STL then it is irreversible back to solid again, similarly like saving Raw into JPG. Art cad users donā€™t see this, but it detrimental to Engineers CAD users.

Also, on a more general note, do you know any one within the ā€œengineering & manufacturing sectorā€ who prefer STL over STEP as an universal communication file? because professionally I couldnā€™t find anyone prefer this.

Professionally, I know quite a few who use SW to create the 3D models first (superior at making geometric object) before importing to 3D Max or Maya for surface treatment like adding texture map, rendering or even animation (as Iā€™ve worked alongside with art department to publish marketing materials, also Iā€™ve used 3D Max before). If there is a choice they wouldnā€™t eve touch STL even if it a go between file i.e. Engineer > Art CAD, before saving into their own native format.

Ask anyone within the ā€œengineering & manufacturing sectorā€ for trying to import Art CAD file > Engineer CAD and they will ask you ā€œwhat are you doingā€? e.g overlaying to measure some stuff perhaps?! I have actually done some thing like this with CMM point cloud data or importing & exporting SAT file (STL like) to determine the robotic arm angles while preventing collision in assembly.

There is a reason why there is a very little depend for Engineer CAD to edit STL, but rather view it only.

Have you actually used any of these programs, like 3DS Max or Solidworks? When you import a file created in 3DS or 3DS Max, into Solidworks, you only get a bunch of meshes without attributes, no smoothing, no curves, NURBS, beziers, nothing. The imported models are for all intents and purposes nothing more than the equivalent of a low res STL. Not useful for editing.

Going the other way, itā€™s the same, except 3DS Max selectively reduces the mesh, but retains the smoothing groups/hints. None of the original features, colors and textures are lost.

No they donā€™t. They contain a mathematical representation of the solid model that is dimensionally the same as the original, but the way it was created is lost. The steps you took to create the object are gone. The program attempts to recognize the features, but the final output is way too different form the original. More often than not, it canā€™t solve how to built it, and simply shows an imported model in the hierarchy. Not very useful either

Dudemaster is right. There are basically two types of CAD file formats, ā€œparametricā€ and ā€œgeometryā€. Parametric files describe volumes and surfaces using mathematical equations, curves are ā€œcontinuousā€ and unfaceted - just a start point, end point, and radius. ā€œGeometryā€ files describe volumes and surfaces as connected facets with vertices, curves are approximated (quantized) and so, faceted.

When I model something in NURBS, I export it as STL to print and set the curve division to get something reasonably smooth without blowing up the size of the file. But I keep the source object in NURBS format for subsequent editing because if you take a parametric object and convert it to a geometric object you canā€™t go back. And if the program doing the conversion doesnā€™t give you the option of ā€œcurve divisionā€ the curves in the converted file could be very blocky.

That being said, earlier this week I watched a video from SIGGRAPH 2017 about a software algorithm that reconstruct a parametric model from a faceted model. So who knows, in the near future maybe this will make it to the public domain.

I think that the next few years are going to bring better availability of point clouds generated by the cameras on peopleā€™s phones. I suspect the tools that Iā€™m going to want will edit that sort of information.

Given that the Formlabs programmers have a finite amount of time, what do we want their priorities to be?
Iā€™d rather have a robust set of automatic tools that create correct supports in more and more cases, and skip tackling different hard problems.

+1 for step file support

From an Engineering & manufacturing perspective, the ā€œmathematical representationā€ you are referring to is and always will be an approximation when saving from Native CAD to STL, which is nowhere good enough for manufacturing, because you cannot measure things in great details in STL and if we canā€™t measure things accurately, we canā€™t modify tolerances for manufacturing.

E.g. it a common practice for mould tool designer to compensate shrinkage depending on injection materials, but sometimes it only shrink in specific orientation. So a simply scaling wont work to a typical ā€œnegative space modelā€ and they also add features in which design engineers misses such as correct split line adding ejction pins and cooling and draft angles including those small fillets to prevent tool wear.

These are the sort of details I know SLA can do and so it would be very wise to use these SLA simulated models to test out the production design along with all the necessary details from a CAD model first before committing to buy 10ā€¦40k injection tooling, which can take anywhere from few weeks to several months.

Those injection tool makers will never accept STL file. This is just an example from one manufacturing sectors, I can say the same thing about CNC press brakes (sheet metal) or tubing, investment caster, forging & CNC machining (still use old G-codes),PCB like Altium start to take in STEP as well even though PCB uses Gerber files for SMT or Wave, even the professional model makers prefer STEP file over STL. eventhough they have tend to have facilities to import STL, because STEP are more flexible.

THAT IS WHY ā€˜ADDITIVE MANUFACTURINGā€™ EXISTED IN THE FIRST PLACE , IT CALLED ā€œRAPID PROTOTYPINGā€ to help out designers to test out the designs before commit to production and we donā€™t have that kind of flexibility if we use STL over STEP.

This is also why you donā€™t see a lot of STL files being shared in the Engineers community like GrabCAD or 3D content Central

In fact, why not ask Formlabs themselves and they will say most of the parts in their machine where manufactured from using a STEP file and not STL.

  • Orange UV cover & tray = injection moulding
  • Resin cartridge = blow moulding
  • Build platform = cast & machine (I think)
  • Wiper blade = injection
  • plug & cable = low pressure moulding
  • Main CPU box = sheet metal
  • Galvo mirror Optics = float glass (method I think)

Iā€™m not here to dismiss the legitimacy of using STL in Preform, but Iā€™m saying it unfair towards Engineers sector who canā€™t use STEP, because currently it is inflexible.

And to answer your question Iā€™ve used many CAD system from both CAD discipline (Art & Engineering) professionally & personally, AutoCAD, MicroStation, 3D Max (Briefly touch it with Maya), ProE 2001 & Wildfire series and SW.

In fact Iā€™m a certified Solidworks professional since 2006 (quite easy actually), but I also graduated and worked in the relevant field of design research & development (this including 3d printing). So yes Iā€™ve used CAD on a daily basis (along with 3D printing) since early 2000 this include communicating with other companies who uses different CAD system to ours.

From my observation with SW, they always add on new features and hope someone will use it (only 15-20% core features), but back in SW 2002 many feature have to be created manually and this simplification features over the years have made some new CAD users complacent in CAD modelling.

I on some rare occasions have to tell how SW support people how to model in SW with some really difficult modelling techniques like dynamic compression springs.

My Channel: (without narration)

But after all these years, I still consider myself as a learner as there are so much out there, because if you ever been to a University with CAD training course either as an artist or engineer, the first thing they will tell you is either Engineer CAD OR Art CAD, but not normally both because it is two very different discipline with CAD in which I so happen to experience both (formally).

I donā€™t consider myself to know everything even within Engineer Sector, but I do know enough that majority of Engineer use STEP and hardly with STL. Art CAD uses STL, which is fine with me because Engineer have different requirement than An Art CAD user.

The file is from GrabCAD (as I cannot show my work online), but it has a 4 window views features typically used in Art CAD packages (Less used in Engineer CAD).