I'm still pretty new to 3D printing so to me it is still magic.  Create something electronically, send it off to a web site and in a week or so I can hold my creation in my hand...that's magic.  This project was to recreate the historic mission in San Luis Obispo.

This summer we had an event at the San Luis Obispo ('SLO' for the in crowd) mission.  This historic building was built in 1772 and I thought it would be a good subject to reproduce in 3D.

My initial target was the mission building, rather than the entire complex.  Modeling and scaling were relatively simple; original (-ish) floor plan sketches and reference photos were available online.   I wanted the finished 3D print to be a Christmas ornament, which wasn't ideal because the mission is rather long and not symmetric.

Since the main entrance, which contains the bell towers, were the most interesting portion, I decided to focus on a roughly cube-shaped boundary which included the entrance.

As I said, the modeling was straight forward.   There were, however, some lessons to be learned before the model was ready to send for 3D printing.

• Obviously textures and bump maps aren't sufficient for 3D printing.  I was surprised at the effort required to get the roof tiles and rock walkways ready to print.  I've written about my approach to the roof tiles.  For the rock walkways, I tried several approaches, including subdividing like crazy (2.6 million vertices) and applying a displacement modifier.  That gave me exactly the look I wanted, but the number of vertices wasn't practical for sending to 3D print.  Eventually, I went with a smooth walkway surface.
• Small pieces are fragile.  I knew that going in, but thought the detail, particularly in the bells, was necessary.  I like the final look, but I expect the bell supports to break early and often.
• I didn't realize large, thick pieces would be a problem.  I modeled the exterior pillars and the remaining interior of the mission as solid pieces.  It turns out that requires a lot of material.   The initial quote for a 6" model was over $5,000!  I reworked the model to create hollow walls, reduced the depth of the base and reduced the thickness of the partial wall to the side of the entrance.  The net result was a price drop of more than 90%.
• 'Designing for Supportability' is a common term in engineering.  A less frequently heard phrase is 'Designing for Paintability'.  I intended to paint the final model.  With that in mind, a smart person would have have modeled the exterior columns and structure, possibly even the roof and supports, as separate pieces.  I am not a smart person.  There were many challanges when it came time to paint, including the fact, as mentioned above, that the bells were very fragile and did not appreciate my attempts to paint them.
• 3D printing requires a 'cleaner' model than rendering.  I discovered several modeling bad habits that have not been an issue for rendering but did require fixing prior to 3D printing--non-manifold edges and faces, intersecting faces and non-quad faces, to name a few.  Relearning a lesson that it would have been faster in the long run if I would have done it right the first time! :)

Despite all the opportunities to excel, I eventually was ready to send for 3D printing.  I have used Sculpteo for the few projects I have attempted.  I am very pleased with the results.  I haven't used any other providers, so I can't make a comparison, but I have nothing but good things to say about the Sculpteo website, process and final product.  They even provide renders of what the finished, printed product will look like.

...and a final render.  It's seems like a waste to only print it. :)