resolution used for the specimen reproduction depend on the scanner and its
proper use, and the skill of those operating the CNC/rapid prototyping machines
(Kadobayashi et al., 2004; Remondino et al., 2005, 2008; Bitelli et al., 2007).
Specimen features measuring less than 1 mm will be lost due to improper scanning
or editing (mainly if the scan left many and large holes to fill in the
post-processing, something that can be avoided by doing more scans from more
different angles). Good scan data can be belittled by improper use of the CNC/rapid
teachers, and museums have different requirements concerning costs and accuracy,
the proper reproduction technique must be chosen depending on their intended
use. For numerous analyzes in functional morphology or anatomy, highly accurate
models are needed, and error margins resulting from inaccurate models have to be
considered. Especially when studying such delicate and complex structures as
diplodocid cervical vertebrae, accuracy should be considered more important than
not only more accurate (0.08 mm compared to 0.1 mm in CNC-milling), but also
better suited for producing models with complex and delicate structures such as
vertebrae or skulls (Figs. 4a and 4b). The problem that the 3-D Z-Corp. printer
at our disposal had a maximum build space of 300 mm x 400 mm x 300 mm, was
circumvented by subdividing the vertebrae into smaller parts. Furthermore, in
the case of the Z-Corp. machine, rapid prototyping machines are much less noisy
and unclean than CNC-milling devices, and the remaining unused powder can be
the CNC-milling process are that the machines do not have such a limited
building space as 3D printers. Especially the more elaborated 6- or 7-axes
devices are thus very well suited for the reproduction of large objects with no
or only minor undercuts (Fig. 4c). Moreover, as this method can be used with
very durable and cheap materials like wood or metal, costs of models produced in
this way are much less fragile and expensive and therefore ideal for educational
use in schools or museums (Deck et al. 2007).
shape of the cervical vertebrae of SMA 0004 required a large number of polygons
to properly describe them. A 3D printer capable of generating slices small
enough to replicate those polygons was therefore preferred in this case study.
Using CNC-milling, the machine output would not have shown all the specimen
features required for further studies.