The Eloquence of Others

Sometimes, other people can state your frame of mind far better than you can yourself.
Take this posting from Tom Madden on the subject of 3D printing;

Here's the deal. We make miniatures, not full size objects. That puts us in the same category as jewelers. The affordable consumer grade 3D printers that everyone is salivating over are not intended for us. They are for the folks who want to make hammer handles and doorknobs,

The technology is advancing very rapidly, but, for now, the equipment and materials that can print objects suitable for use as-is cannot render the level of detail we require, and the equipment/material combinations that CAN render a suitable level of detail print objects that are very expensive and/or aren't suitable for use as is. (Brittle; dimensionally unstable.)

I've been working in this field since 1995 and am very impressed with how rapidly things are moving. I'm also a resin caster so I look at everything I make as a potential pattern. That lets me justify spending a bit more than if I was after something I only needed one of. But the cost can still make you gulp. This photo shows portions of cast resin HO heavyweight Pullman sides with the patterns made by two different 3D printing technologies:

http://www.pullmanproject.com/SideBoth.jpg

The pattern for the upper side was made by multijet printing (MJM), the lower by stereolithography (SLA). The MJM pattern was done by moddler.com using the same material and equipment as Shapeways' "Frosted Ultra Detail" (FUD) but with the customer able to specify build orientation. The SLA part was done in house. The surface of the MJM part is a bit rougher than the SLA part but is still acceptable. Both are from sets of four half sides. There are newer MJM machines that can build complete HO Pullman sides at that resolution, but the highest resolution SLA machines are limited to a 5" x 5" area. (That's why I wanted to try the MJM technology.)

Now the "gulp" part. The set of four MJM half sides from moddler.com cost $250, the SLA part cost me $150, less than half what an outside customer would have been charged. Shapeways quoted $84 to do it in FUD, but they would have built the parts standing on edge rather than face up. That is a fundamental problem with Shapeways and is non negotiable. They load as many parts as possible on a build platform, placing each to minimize its footprint. If the detail face is built as a side wall, the detail on it will be degraded. I use Shapeways a lot and can work around that for many of my projects, but not for car sides.

The real problem is, the companies that can print objects with the precision and level of detail we want are focusing on commercial customers with deeper pockets. Plus they are using $six figure machines. Competition and demand should bring prices down, but at 77 I don't expect to see a 3D printer capable of cranking out ready to use parts like those sides sitting in my workshop during my lifetime.

Tom Madden

Tom has eloquently stated what I have believed about 3D printing and it's potential and limitations far better than I ever could.
And I will happily admit that we are developing some freight car parts using rapid prototyping. Without the current technology it is doubtful that we could have satisfactorily produced the parts we are trying to create. But we are only using the technology to help us develop the patterns. Production runs will be traditionally cast and the final "print" for the pattern will indeed cost us 3-4 times as much as the test prints from Shapeways. 
3D printing is a tool, not an end unto itself.