(or “Why can’t we just 3D scan a ring and make copies on a 3D printer?”)
Last Updated 14 June 2016.
This is a question which seems to come up mainly among advanced students or those already experienced with using CAD for jewellery, but I do get the occasional person who has just been to a trade show and seen the latest 3D scanning machine released onto the market. Either way I reckon this is a good question to answer next, as it ties in with several other FAQ articles I’ve written previously for this site.
It seems for the past 10 years 3D scanners have been improving rapidly in terms of scanning resolution, scanned mesh quality, assembly of data, ease of use, and efficiency of file sizes (otherwise known as decimation to those who work with the machines).
You’ll even see salesmen and technicians at trade shows demonstrating how easy it is to 3D scan a ceramic or plastic part or a visitor’s face, import it straight into their program, and manipulate the mesh form in full colour.
So why doesn’t this scanning device work the same way with jewellery?
Well, it kind of does, but not at all in the way you would expect, and as of the date this article was written 3D scanning still faces some serious technical limitations. To answer this question fully, I’ll have to answer the question in three parts: the differences between the various CAD geometry types used by 3D modelling software packages, the problems with model conversion, and the technical limitations of the 3D scanners themselves.
Different CAD Programs Mean Different CAD Geometry Types
CAD software developers have come up with many different approaches to creating a 3D object, all based upon different CAD geometry types. Part of this stems from the fact that CAD is used in so many different ways for so many different industries, but part of it is down to many different software developers each arriving at different solutions to the same problems.
- NURBS surfaces (such as Rhino or 3D Studio Max)
- Subdivision surfaces (such as Maya or T-Splines)
- Polygon Meshes (This is the geometry type what we get from a 3D scanner. Point clouds and solid modelling usually fall under here as well. In jewellery, ArtCAM and ZBrush would be examples of mesh modellers.)
Parametric modelling and Direct 3D modelling would also fall within these categories, usually as a variation of either NURBS surfaces or Solid modelling.
The reason why all of these different CAD geometry types exist is because each one has its specific areas of efficiency and specific purposes. I’ve made a separate article dedicated to just discussing the advantages of different CAD software packages.
The problem with CAD model conversion is these three different geometry types are all mutually exclusive, and we start running into all sorts of difficulties when we try to converting back and forth between the three various geometry types.
Retropology and Reverse Engineering in Jewellery CAD – Converting Between Different 3D CAD Model Types
The most commonly requested conversion has always been from NURBS or subdivision into meshes, mostly because meshes are the simplest format in which a 3D object can be stored as well as the standard file format for Rapid Prototyping machines. Over the decades countless conversion algorithms have been developed to make this one-way conversion efficient, even if no conversion ever gave 100% precision.
Going the other way from mesh back to the other model types, however, is much harder. This is partially due to meshes not really recognising edges of objects the same way as NURBS surfaces do, and partially due to the fact that mesh faces cannot be anything other than flat facets.
Steps have been taken by various developers to make this conversion easier within software. The whole field of reverse engineering developed around the prospect of having to convert meshes back into NURBS. Many years were wasted on finding an efficient way of making this conversion. That all changed a few years ago when the developers of subdivisional modelling tools came up with something called Retropology.
Retropology tools are designed to help map subdivision surfaces to mesh skins by quite simply letting designers apply the new flexible surface facets one by one onto the mesh skin. Programs like T-Splines and Zbrush have since included retropology tools in their software to allow the manual creation of subdivision surfaces based on a mesh which can then be converted into a more easily editable form.
The ultimate goal of these retropology tools (and part of the reason for their development) was to make it possible to take a 3D scanned mesh and convert it cleanly into a workable surface in NURBS. However, even with these tools it’s still not quite that easy.
Limitations With 3D Scanning Hardware
This brings me to the limitations of 3D scanning hardware.
In spite of all the progress made on 3D scanners, three major problems with 3D scanners remain before they become useful for jewellery CAD:
- First, getting the resolution down small enough so that we can read clean and delicate details on something as small as a ring is a surprisingly big ask, even for some of the most powerful 3D scanners on the market.
- Second, scanners don’t deal with reflective surfaces that well. In order to read a surface, the object must be matte. This requires coating the object with a coat of white paint. While this isn’t a big deal on larger objects, the white paint can obscure or clog up small inset details, and while it doesn’t damage the surface, it could even prove difficult to wash out again.
- Third, the 3D scanner needs to be able to reach all the angles of the surface, or guess at what’s missing. If the scanner cannot reach a specific angle, that area will be left as empty space. In practice this leads to meshes with gaping voids on one side or another, which can be time-consuming to repair even with sophisticated tools.
The result of all three of these problems together is that we tend to end up with 3D scanned ring objects which can often not have the resolution or model integrity to be usable even for retropology, let alone 3D printing.
So Why Bother With 3D Scanning For Jewellery Then?
As it happens, even with the limitations we’re currently facing, there are still many good uses for 3D scanners, such as:
- Scanning larger objects and reducing them down in size to save us a significant amount of modelling time.
- Scanning an existing piece of jewellery to use as the starting point for creating another piece of jewellery, allowing you to reengineer an existing piece without taking the other piece apart.
- Scanning surfaces which would be too difficult to measure to use as the starting point for making an object. One good example would be the service offered by Alba Rose jewellery where they 3D scan engagement rings and create bespoke side bands.
Even if we cannot yet efficiently use 3D scanned data for 3D printing as it is, we can still use that data as the start for building another design.
Currently Available 3D Jewellery Scanners
With that in mind, I present a list of 3D scanners I’ve confirmed as being designed specifically for scanning jewellery-sized objects:
- eQUALITY tech Rexcan series Jewellery and Dental Scanner (the DS3 and the previous DS2 model in particular seem popular)
- Roland LPX-60 Scanner
- Photon Matterform Portable 3D Scanner
In addition to the above, there are several other scanners which claim to have the resolution to be able to work on jewellery size objects, and I’ve also seen being used for that purpose as well:
I’m sure this list will only keep growing.
I’d like to include a couple of disclaimers for the list above:
- The main reason why I’m listing them all here is because it’s hard to find this information all in one place anywhere else. I have not tested these machines thoroughly, so I cannot as of yet give any official advice as to which ones are better than the others.
- As 3D scanning technology is slowly but inevitably improving, and it seems every year the technology has made another series of big improvements, this list above will likely be subject to frequent and rapid change.
It seems so many engineers see so much promise in 3D scanning that it will surely continue to develop rapidly, and I suspect it is only a matter of time before what I’ve written here will need to be updated again as the technology matures. In the past two years 3D scanning technology has been refined to the point where it has become affordable enough for general small scale use by jewellers or hobbyists, and we’re already starting to see desktop 3D scanners become cheap enough for hobbyists as well. Who knows what we’ll see in another two years?