Monday, January 7, 2013

WAGR R Class Diesel using 3D Printing

The R class has always been one of my favourite locos and just fits into the 1960's timeframe of my layout.

Over the last few months, I have been working on a model of the R class using 3D printing techniques.  While the technology isn't quite there yet at delivering a finish comparable to injection moulding at an affordable price, it does appear to be practical to produce a complete loco using 3D printing, if one is prepared to put up with some imperfections.

No doubt, the technology behind 3D printing will continually improve, so my strategy is to accept the current level of quality as part of the learning process to build a "proof of concept" model.  In a few years time when the quality has improved, I can get new "prints" done and get a new, better model for minimal extra effort.

Today, I received my first complete body for the R class as a 3D print from a company in Belgium called i.Materialise.


The major deficiency of the current (affordable) 3D printing processes is that the resolution is not fine enough to get a nice, smooth surface finish.  The laying of the material as the model is being "built" results in a roughened "striated" surface and also a "stepping" effect on curved surfaces, particularly, in the case of the R class, on the roof.  Whether the quality of surface finish is acceptable depends very much in the individual modeller.

Note the imperfections on the large square radiator intake grille in the form very fine "whiskers". These are due to a minor problem in the design which will be rectified in subsequent prints.


The beauty of the 3D printing process is that virtually all the details can be "printed" as part of the process, providing the details are, of course, built into the 3D artwork in the first place.  In the case of the R, it is really only handrails and horns which will have to be added to complete the body detail. I decided to built the raised numbers on the cab side into the 3D design and they appear to have worked fairly well.


This view of the side detail shows the striated effect visible on vertical surfaces.  While it would be possible to sand the flat surfaces to get a smoother finish, it would be difficult to do so without damaging the fine details.


This top view shows the stepping effect visible on the curved roof.  As my layout is close to "eye level", this effect will not be as obvious when the loco is operating.


I already have a working chassis using a custom-made power bogie from Hollywood Foundry and matching unpowered bogie. The following photo shows the "loco" on the Hollywood Foundry bogies, with very temporary underframe and temporary weighted "fuel tank".



I am currently working on the 3D design for the bogie sideframes and plan to also do the underframe and fuel tank as 3D "prints".  The underframe will be designed to take 2 powered bogies, if desired.

Although I have no intention to produce the R class as a kit, there is the option of making the 3D designs "public" through the i.Materialise website so that anyone interested in building an R class can order "prints" for themselves of the major components.  I can easily make available the ordering details for the power bogie, etc. from Hollywood Foundry, perhaps as an addendum to this blog.  As an indication, the cost from i.Materialise for the 1-piece body as pictured above is $100 + $30 shipping per order.  Delivery time is about 2 weeks from ordering.

I don't have any intention to do the design for an RA.  There are some major differences between the R and RA, including the RA being longer overall.  I estimate it would take about 30-50% additional design time to modify the R design to produce an RA.

14 comments:

  1. thats a very nice job i was wondering how long before a 3d printer was used

    ReplyDelete
  2. 3d printing is a very real option for producing models, if only as masters for urethane kit production. as noted the difficulty is cost. Fineline in the USA does good work with a finish suitable for modelling use, as does Mark 4 Design in NZ. One limiting factor to keep in mind can be build size, some machines and process are limited to a maximum size of about 100mm cubed....

    ReplyDelete
  3. Another factor which is an "unknown" at present, is the long-term stability of 3D prints when used for actual models. Mark4 Design do not recommend use of their 3D prints for actual models in scales larger than N, instead suggesting their use as masters for casting. A major downside of using 3D prints as patterns is that this imposes many more constraints on the design to ensure that it can be cast successfully. On the other hand, "pure" 3D printing can print shapes which would be impossible to cast, e.g. with complex overhangs, etc.
    The particular material I am using was chosen based on the combination of sufficient build size to allow 1-piece construction, reasonable resolution and surface finish, the ability to take normal hobby paints, and temperature stability to approx. 60C, plus reasonable cost. (Some materials can lose their integrity as low as 43C.) I ordered test prints of the short hood only from 3 different suppliers in 5 different materials, before deciding on Prime Grey for the complete body and other major parts which are in progress.
    The Prime Grey material has only been available for about 18 months so there is no real experience on how it holds up long-term.

    ReplyDelete
  4. I would get one of these in N Scale if they could be printed on demand from some 3D printing company.

    ReplyDelete
  5. Beccas
    If you are serious about an N scale model, it would be possible from i.Materialise but I couldn't guarantee the suitability of the finished product, as I have designed the loco body based on S scale. For example, some small details may not print properly.
    In N scale, the smaller build size would suit a higher-detail material such as i.Materialise High Detail Resin and the scaling is possible at the time of ordering, e.g. 1:64 scale x 40% = 1:160 scale (N scale). The cost for one body would be approx. $70 plus shipping from i.Materialise.
    I don't intend to make the loco body available for others to order for several months at least, until I complete and test the other parts of the loco. But, if you really wanted one in N scale, it would be possible.
    Another possibility would be for you to order from Mark4Design in NewZealand which are not "print on demand" so you would have to do some "leg work" to organise the printing, based on my STL file. They do very detailed work including N scale locos and wagons, as illustrated on their web site (http://www.mark4design.com/).
    As a "favour" for another modeller, I have already organised printing of a modified version of the R class bogie sideframe in 10mm/ft scale (just over twice S scale), although going to a larger scale has less potential pitfalls than going to a smaller scale.

    ReplyDelete
  6. Regarding the marks on the curves from the printing process, would it be possible to use putty to fill and sand back slightly, just to take the noticable edges off at the least?

    Apart from that, definitely worth the effort - I've been looking at 3D printers on eBay, and they aren't that expensive, but certainly not as good a resolution.

    ReplyDelete
  7. I would be keen to try printing this on my RepRap, in N scale - hopefully I am going to be able to get some really good detail when I get going... I think it could be easily adapted to be one of the current NZ diesel locos...

    Would you be able to share your files?

    Nate

    ReplyDelete
  8. Hello Kiwinet
    Sorry, but I am not willing to share my files. I had some test parts printed on a extrusion type printer in S scale and the results weren't all that great, so I don't think the quality would be very good at all in N scale.
    As far as I am aware, most NZ locos are EMD or GE and the only EE units had full width cabs, so no resemblance to R class.

    ReplyDelete
  9. Came back a bit late. Sorry about that. $70.00 is bit too much for me. I will keep watching and hoping 3D printing will get cheaper.

    ReplyDelete
  10. Hi Richard.
    Perhaps a few readers might like to know how you started out approaching the cad/cam computer system. Which one you used and why. Might make a great article for those wanting to make a start in computer design. For example what did you do to learn to use the software. Read a book? The manual played with it or tried building small objects

    ReplyDelete
  11. Hello Peter
    I have been using 2D CAD for many years, and dabbled in 3D CAD about 10 years ago for designing a house extension using special purpose home-design software.
    For this project, I started out with free software from Autodesk called 123D, and I first started on a very simple shape, a wheelset, in fact. I thought I might be able to get accurate wheelsets 3D printed to use as wagon loads. I found the user interface with 123D not very intuitive. They had tried to make it clever and "different", but it was hard to control it to do what I wanted in any predictable way.
    I did start part of the R Class using 123D but it was fairly hard going.
    Based on a suggestion from another of the S scale group, I trued a free demo copy of Rhino Version 4. I found that Rhino was much easier to learn and worked similarly to 2D CAD which I was used to (Autosketch). Within a couple of hours, I had recreated the R Class short hood which I had started in 123D, and it seemed a whole lot easier all round.
    I didn't read a book or go through any structured training. I just started off drawing basic shapes and sticking them together. I had some pointers from my brother who does 3D CAD as his profession, for example about drawing complex outline shapes in 2D and then extruding or rotating them to produce a 3D shape. If I got stuck at any point, I read the online help or searched for information on the Internet. There are lots of tutorials and short videos available. Unfortunately, some of the people who produce them don't really understand what they are doing, so you have to discard the rubbish.
    Rhino 3D is not perfect (and not free), but I have found it pretty easy to learn and it (mostly) does what I want it to do.

    ReplyDelete
  12. if this could be printed in 3mm scale (1:101.6 or British TT) i would order one just to test it. yes, the stability of material is an issue ... i've ordered a few models from shapeways in FUD and i think their overall quality is at the low end. i've heard one modeller did some stuff for puffing billy in G scale for outdoors ... allegedly the material started to deteriorate quickly in the australian sun.

    ReplyDelete
  13. You have done a very fine job with the virtual model. Unfortunately, the iMaterialise print just does not cut it compared to a state of the art RTR model from almost any manufacturer.

    For me, your work is a very inspiring "proof of concept", and I am eagerly waiting for the day when 3D prints available for hobbyist fulfills its promises.

    ReplyDelete
  14. I suggest that using the digital model and various type of 3D printer from many companies as Sunruy 3D printer technologies, 3D printer deposits microscopically skinny layers of the material, and also the print step by step materializes because the layers square measure designed up step by step. The quantity of detail potential during a 3D print is decided by the thinness of the layers, and also the material is something from resin, to ceramic powder, metal, or perhaps glass. Click here Best 3D printer manufacturer company in China.

    ReplyDelete