3D Printing an Autodesk Revit model

2,136 views
Skip to first unread message

Chris Pittman

unread,
Dec 12, 2012, 11:50:55 AM12/12/12
to make...@googlegroups.com
Hi everyone - I recently joined this forum after ordering a Rep 2 in October.  Now I'm still eagerly awaiting it to arrive (should ship this week) but in the anticipate I've begun brainstorming more and more stuff I'd like to try printing.  By profession I'm an architect and currently all of our designs are done in Autodesk Revit.  I figure the process to go from Revit to a format I can print won't be cut/dry, but I was hoping someone here might have some tips.

I can simplify my Revit model and only export the big pieces (walls, roofs, floors) and leave the extraneous stuff out.  And I can export to various formats, dwg, dxf, fbx, sat, and ifc.  

I have access to just about any design software needed (especially anything from Autodesk).  

Does anyone have any ideas of best method?  Or am I better off taking the conceptual design - remodeling it in another piece of software - then printing it?

Any help is much appreciate - thanks

An image attached of a section of the building I'd like to attempt printing.

Jetguy

unread,
Dec 12, 2012, 12:57:56 PM12/12/12
to MakerBot Operators
Holy overhangs Batman.

Um, you might want to study FDM printing that we do a little more
before tackling that one. Every single part represents a challenge,
especially the odd I beam structure. There's only a ccouple of
orientation that make sense and every single part of that structure
must be separated.

I mean you could print with full support and waste a spool or two of
plastic and then have to remove all that waste for a crappy final
product at best. Please, don't even try this with plastic hovering at
$20 or more per pound.

http://www.makerbot.com/blog/tag/design/

http://wiki.makerbot.com/design-tips

http://wiki.makerbot.com/design-guidelines-for-makerbots

Also note, you can only print STLs as a file type.
> <https://lh6.googleusercontent.com/-FKrZXiARHxY/UMi12SdUXpI/AAAAAAAAEQ...>

Jetguy

unread,
Dec 12, 2012, 1:14:12 PM12/12/12
to MakerBot Operators
Sorry, and I wasn't trying to discourage you. It's just you presented
quite a challenge with the question and specific example, along with
the lack of understanding of what files types are supported. I want to
help you as best as I can but you need a crash course in printable
modelling to understand the limitations of the machine.

And to that end, guys are having trouble with Replicator-2s printing
well designed large models with the extruder. Your challenge would
certainly result in a lot of expensive wasted plastic. I'm being harsh
now, to set your expectations so that when you get the machine, you
don't assume it's some problem with the machine and not the real
fault, your model. The truth is, not everything is viable to be 3D
printed using FDM. It works great for much our items and we work
around the limitations. The large flat surfaces of the wall and such
can be printed, but likely are never as good a original sheet material
such as a laser cut model. Again, a building like this represents a
huge challenge. Not exactly a good first project to jump into.

I'm sure there are some folks here who can further help guide you on
how to break that up and orient the parts to be printable but again,
that's a huge learning curve.

I say it would be best to read the sources I linked, and start going
around Thingiverse and seeing models that people have printed. By that
I mean there are lots of example STLs but make sure you see pictures
of an actual printed one on the same page. If nobody has printed it,
it might be cool to lack at in CAD, but completely not feasible on a
Makerbot. Again, I want to be clear you can print nearly any object
with support. The huge question is how hard is that wasted pastic to
remove later and reveal the part you want inside an further at what
cost. The other killer is that many people are having extruder issues
on long prints. Obviously printing with support means more plastic per
layer and now insanely long prints that take several hours and maybe a
whole day. How upset would you be at 10 hours into a print job and a
roll of plastic to have a ruined print?
> > <https://lh6.googleusercontent.com/-FKrZXiARHxY/UMi12SdUXpI/AAAAAAAAEQ...>- Hide quoted text -
>
> - Show quoted text -

Sean Tu

unread,
Dec 12, 2012, 1:18:43 PM12/12/12
to make...@googlegroups.com
Pretty sure the OP understands that he'll have to print the thing out as component pieces, ie, walls, roof, floors, and then assemble the thing or be wasting a lot of time and plastic printing the supports.

Looks like there's a POC project for exporting STL files with Revit

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
To post to this group, send email to make...@googlegroups.com.
To unsubscribe from this group, send email to makerbot+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/makerbot?hl=en.


Chris Pittman

unread,
Dec 12, 2012, 1:19:35 PM12/12/12
to make...@googlegroups.com
No discouragement taken - I realize going into this that printing one of our architectural models wouldn't be easy or cut/dry.  I probably didn't make it clear enough, but I have big doubts I can take anything directly from Revit and print.  I figure more than likely I'll need to take the conceptual shape and remodel that in a fashion more fitting with FDM printing, removing the extraneous stuff from the interior of the building and getting it down to the basics to represent our idea.

Originally I bought the Rep2 simply because its technology I'm interested in and at heart always been a tinkerer - it wasn't until after the order was placed that I was approached about the possibility of printing our designs.

Thanks again

Chris Pittman

unread,
Dec 12, 2012, 1:29:16 PM12/12/12
to make...@googlegroups.com
Wow - thanks for the link - that comes in handy, lets me skip a few steps right now.

Thanks

Shawn

unread,
Dec 12, 2012, 1:33:33 PM12/12/12
to make...@googlegroups.com
As is, this would be challenging to say the least. BUT, I can see some
simple separations that might make it easier.

Flip that model onto it's back, with the open end pointing up. Print
the door section separately - possibly with the roof as a separate
piece. Print the interior ceiling as a separate model. The end caps or
I-beams may need to be done on their own as well and attached later.

The trick is to avoid large overhangs, or unsupported sections.
Otherwise you need a lot of supports generated. (Just printed the
Ironman Mask - about 1/4 spool for supports alone... sighs).

As Jetguy mentioned, some models are not well suited for 3D Printing.
BUT, if you can slice em up and assemble things later, I think just
about every model can be made to work.

As for file formats... you say you have access to most Autodesk
software. Autocad, 123D Design, etc. support export to STL. If you can
get your model into there, you should be all set. (I don't work with
Autodesk products very much, so don't know all the
capabilities/limitations....)

All this ignores the issues Jetguy mentioned regarding long prints on
the Rep2...

HTH

Shawn
> <https://lh6.googleusercontent.com/-FKrZXiARHxY/UMi12SdUXpI/AAAAAAAAEQI/v9khu_IVgVw/s1600/Untitled-1.jpg>
>
> --
> You received this message because you are subscribed to the Google
> Groups "MakerBot Operators" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/makerbot/-/XkBdADIJhG0J.

PropellerScience

unread,
Dec 12, 2012, 1:55:47 PM12/12/12
to MakerBot Operators
You could try printing that all at once with full supports turned on.
I did that with a cubicle design I did at work when I first got my
Replicator. I thought it would turn out a blob, but I was surprised
how well it worked. It also gave me an excellent idea of what didn't
work, and why. So give it a try!

funmakerBart

unread,
Dec 12, 2012, 1:58:27 PM12/12/12
to make...@googlegroups.com
I'm also an architect, ArchiCad, bought the Rep1 in januari. Took me an half year to print decent dual colored with acceptable speed and finish. After that it turned out it was a big mistake to think that CAD 3D software was suitable to produce watertight stl's. The good thing: I know now a dozen more programs more than earlier regarding conversion, clean up of stl's as well 3D design modellers.
And it turns out as well, that there is a big difference between the slicers , how the handle non-manifold stl's!
I know now that there is not one solution / workflow from CAD file to 3D model. I choose every time a different combination of modeller, cleaner, slicer.

But to not disappoint you, for a first good result: first print in one color, don't make a model to print at once. Roof and ceilings / stories / facades have to be printed separately. 3 seconds glue is your friend.. I managed sometimes with good tweaking to print at once, but it's too time consuming regarding the result. So: separate parts by exporting them separately  And think as you did when you modelled with cardboard at the university when you separate the model parts.
Regarding CAD and not producing too much slice errors: make it so that all your elements (so floor / walls / columns) have a space of 1mm to each other. This makes it possible for the slicer to know to which object the triangles belong of the stl. Otherwise the triangles of the bottom of the wall are on the same level as the floor resulting in a lot of errors. Slic3d gives the best results for this specific problem, than RepG, than Kisslicer.
Also: spend time to calibrate the printer and have basic working settings ready. 

Have fun!
Bart



Op woensdag 12 december 2012 17:50:55 UTC+1 schreef Chris Pittman het volgende:

Chris Pittman

unread,
Dec 12, 2012, 2:41:51 PM12/12/12
to make...@googlegroups.com
Thanks everyone - great advice - can't wait to start experiementing

funmakerBart

unread,
Dec 12, 2012, 2:50:36 PM12/12/12
to make...@googlegroups.com
So quick guide:
- measure your fillament, choose ABS or PLA, decide how you to stick it to the platform (Hairspray, tape, Kapton), reduce oozing, 
- calibrate your printer with the callibration cube, know the speed limits as well acceleration settings. 
- replace the plunger design of Makerbot with the MK8 mod of emett, 
- make an enclosure for the printer to prevent warping of ABS
- for PLA cooling is important by slowing down or an additional fan. 
- start with a model with one color.
- make first simple models with only a slab, walls and columns in different sizes, so you know the limits for the desired scale of the print
- be sure all the solids of the CAD model don't share a surface and edit the model that there is everywhere a gap of 1mm
      (I tried as well to model the building again in another modeller, just for stl output. Didn't work well, also real CAD software is so easy when you are used to it)
- break up your model in easy to print parts, best so that it won't need support
- export to 3ds works the best for me
- import the 3ds files in Netfabb. Scale them to your needs, move them to the origin xy = zero. Be sure the lowest part of each object is at Z=0
- repair them in netfabb with the red cross buton
- after that export them to ASCII stls
- combine the stl's in Makerware, Repetier Host, Slic3d, or Kisslicer pro to fill your platter.
- print and watch
- tweak and remember settings and probably repeat above
- love and hate your decision to print3D

when going dual printing:
- calibrate the offset of the extruders as well tweak the diameter/flow that the extruders produce about the same quality / layers
- print at the lowest temp to prevent oozing
- In the CAD software export the objects, but be sure that they are associated to the same origin
- in netfabb, select the two 3ds files together, and manipulate them together (scale, move), so they keep the right relation to each other (they will look as the end model together)
- in ReplicatorG you can merge the stl's, but you can't use support
- In Kisslicer you can't yet (next version) combine stl's but you can print supprt with the other extruder
- with slic3r you can combine multi stl's as well generate support. Also it slices CAD based stl's better imo.
- print with RepG

Bart


Op woensdag 12 december 2012 19:58:27 UTC+1 schreef funmakerBart het volgende:

Eighty

unread,
Dec 12, 2012, 5:59:04 PM12/12/12
to make...@googlegroups.com
On the subject of Revit conversions...

I prefer solid modeling, as opposed to meshes. I was pretty disappointed when I exported a family from Revit to dxf, and it was all meshes. So now there would be a lot of cleanup and babysitting. For now, I've been sticking to solid modeling in Autocad.
I'll have to check out the link above, and see if someone has made it easy.

whpthomas

unread,
Dec 12, 2012, 11:36:35 PM12/12/12
to make...@googlegroups.com
Hi Chris, if you only want a small scale model (ie the size of the build platform) the Rep2 can span large gaps quite well. Here is an example calibration print I did with the medium setting, and I am in the process of doing a HQ one at 0.1mm layer height and the results so far look even better. The PLA seems to bridge gaps with only a small amount of sag. My new Replicator 2 has successfully bridged gaps of 90mm with a 0.2mm layer height - this is with a stock Rep 2 using Makerware. You can trim any hanging strands with an exacto knife if need be once its done.
230c-calibration-1.jpg

tvett05

unread,
Dec 13, 2012, 9:32:26 AM12/13/12
to make...@googlegroups.com
Myself and a couple other buddies are 6th year architecture grad students who purchased a Rep 1 to print for ourselves and other students.  We recently landed a job a with a local architecture firm who wanted a small hand full of massing models at 1" = 50', and a single larger one at 1/16" = 1'.  The source model was in Revit. 

Our work flow was to export only walls, floors, roofs, windows, and doors as a .dwg into Rhinoceros 5.0.  The .dwg will be completely useless for making a solid model, and will export as a mesh in Rhino which will require the commands meshtonurbs and mergeallfaces to make modeling in Rhino easier.  Next we deleted the original mesh, exploded every new polysurface and basically removed everything but the exterior wall face and joined every single surface together until it was a single closed surface.  It was a pain in the ass and took about six hours.  For the large model we added thickness to the exterior wall, put in a few of the interior walls, punched out windows and made the roof a separate print.  The small prints were easy, took about 1.5 hours each.  The large print took 18 hours, and ended up splitting at the last inch, making the model unusable.  See attachments:      

FunmakerBart is absolutely correct that for a model like yours all components will have to be done as separate pieces.  We have found that massing models are the best use of the printer architecturally.  Sometimes a good ol' Exacto knife or a laser cutter with sheet material will be best suited for the job.       






Reply all
Reply to author
Forward
0 new messages