What 3D Printing Really Looks Like

I model in O scale as we are building a 20x40 triple deck ON30 railroad.  For our needs we are NOT looking for closeup diorama detail, we are filling scenes (large scenes) with things that have the general appearance of the objects.  I priced out some of the details we needed and we quickly exceeded the price of a 3d printer.  So we got a 3d printer.  For our software we use a terrific little $49 piece of software from Cubify called Invent.  You don't draw anything, you build things.  It is very quick and everything I have designed with it 3d printed.  This is a major issue with most cad related softwares, it doesn't necessarily 3d print well.  Like I said I love my low res 3d printer, it works for our needs.  Russ

stationary boiler for a machine shop power plant (Cubify Invent screen)

3d printed as one piece and it is hollow

o scale hollow pipe fittings

oil dock at Chama

internal goodies for a machine shop, not yet painted or detailed

some more goodies

an On30 gon printed as one piece ready fro trucks and couplers

an air pump that is going on a mow rail car

John - the 800 lb elephant, indeed. A 3D printer could make a nice to very nice rendition of a badly done data file, all day long. Should I bring up the ancient computer programming acronym GIGO? (I am an ancient retired computer programmer - mainframe, that is.)

===

Why all the contention? Can't a new technology (or any idea) come along and be an improvement on or help with what already exists, rather that assuming that it will or needs to be a replacement for it?

For the anywhere near future, 3D printing will be mostly a wonderful prototyping help, and the tool-and-die maker's best friend. It is not a production device except for items that are either low-end (cheap printers), very expensive (expensive printers) or not production-rate time-sensitive.

Injection molding from 3D printer-made prototypes/tooling is a good thing; replacing one with the other - broadly - is not.

And, when you need a zinc or brass part, I can't see 3D "printing" of molten metals.

It's a great advance, but just part of the story, and not worth arguing (as opposed to debating) about.

======

Personally, purely personally, I, for some reason, do not find the subject of interest. It's a good thing, but a bit of a yawner for me. Go figure.  

I'm most interested in the solution of making the 3D model description.  If that was possible without a lot of labor and a steep learning curve, I could live with the limitations of 3D printing for a lot of projects.  Generating the data file for a complicated part is still going to take a lot of CAD time and experience, I don't see any way around that right now. 

I've been hearing about 3D laser scanners that will allow you to duplicate a part, but I figure if the part is already made, I can likely buy it as a rule, so I don't know how useful that might be.

If you make the 3d model there are places referenced earlier in the thread that can print them at quite high resolution. if one were to make a model with details slightly larger that could be sanded and filed down I would think the end result would be quite nice. For example making the steps slightly thicker so you could finish the surface

Russ, I do like your applications. Something nobody would notice has the rough edges. If you keep that at a fixed distance form the eye that issue goes away

How do you "redraw" it and have the correct output files to print the 3D part?

It's not finding the software to do the design, it's the steep learning curve to do anything other than draw a cube or cylinder with those software packages!

Im with John, you learn how to use the software which takes quite sometime to do proficiently and then you get caught up in other things and when you come back you have to re-learn how to use the software again. Its not that simple, just like using the layout planning software is not that simple. I can be simple but you get yourself caught up in trying to figure out how to do some mundane task that would take 2 seconds on paper and after 2 hours you have to quit. Stupid computer stuff!!!!

Thank you for making my point.

I have been using CAD for a living since 1999

Given the same experience level, I wouldn't have any qualms about doing this job either!

Can I interest you in what I'd consider a simple project, say programming and board design for a light controller? See my point? You have to deal with the learning curve.

I mentor at the local jr high and my 6th and 7th graders are getting reasonable models done withing a couple of weeks. That is drawing out a design on paper, figuring out how to get that into the computer, then 3d printing the model.  I have 20 students in one class and 22 in the other.  Only about 4 or 5 dont get it and are having troubles just getting their idea down on paper.  So they are getting art lessons this semister and will be back at it next semister. For our class room environment we use Cubify Invent and MakerBot printers.

If 3d printing is difficult how about wireing your layout?  I have built model railroads for folks for about 50 years now and have lived through ac, dc, dcc, dcs, tmcc, astrac, and lots more.  So all things being relative 3d printing is as hard as . . . . . .

Almost all of the 3d software for 3d printing have free trial downloads.  Try one and see how hard it is before you make comments like STEEP learning curves.  Lots of folks are having good success with the freeware Sketchup and a plugin for outputting 3d printable files.  There also some really great 3d forums just like this one for getting answers to your questions, some are software specific and some are hardware 3d printer specific.  All one has to do is go see whats available, watch a few of their tutorials, and try their trial software.  If its to hard to use or understand, look at another til you find one that seems like it will work for you.  You can tell how good or bad the software might be by visiting their forums.  Anyhow, like have said I love my low res 3d printer.  Russ

"Maybe we hobbyists could share these files once they are created" 

https://www.thingiverse.com/

I'll try to answer these questions as best I can.  The caveat applies that I know Autocad pretty well, I have some experience with 3d printing, but have never directly utilized 3d scans with a 3d modeling software.

A 3d scan could be made of most any part you desire to copy.  Big or small.  You could 3d scan a prototype locomotive.  Or a room and it's contents.  Or something like a 1/48 loco pilot.  And the model pilot should be as simple as finding a shop that has a 3d scanner.

The results of the scan - often known as a point cloud - can then be imported into any number of programs.  The latest version of Autocad supports point clouds.  I would presume that SolidWorks and Inventor (the Autodesk equivalent of SolidWorks) would also support this technology.  From a quick reading, Autocad can interpret planes, edges, centerlines, and other obvious geometry found in the scan file.  While the imported file is not specifically a .dwg file that can be readily edited in Autocad, it can sort of be traced to create entities in the drawing file that can be edited.  It sounds like planar geometry would be simple to convert, while more organic shapes (think f7 nose) would require more effort to convert to drawing format.

Autodesk point cloud use in Autocad

Now whether this would be an effective way to create the 3d file for your specific pilot or any other part depends on several factors.  But the simpler the part, the more effective it would be to just model it from scratch (presuming access and skills with an appropriate program).  It all really depends on complexity of part and quality of data available to model from.  Accurate, comprehensive dimensions make drawing even more complex objects pretty simple.  The head-scratching over inconsistencies is what takes time and affects part quality and accuracy.  The bottom line is that if you had the part you wanted to copy and modify readily available, I would probably prefer you just give me that to work with rather than the 3d scan.

Regarding 3d printing and scanning providers, the answer is again yes.  At least for printing, and I suspect there are now plenty of shops that can scan things too.  

The obvious big player for 3d printing for a lot of people is Shapeways.  They have lots of media available, with lots of corresponding issues, costs, and results.  But an alternative I have used several times is to search locally for 3d printers.  A simple google search will yield results, or you can go to a site like 3D Hubs.  You can find local "hubs" offering printing services based on your zip code.  There are varying levels of print types and quality, materials, and costs.  Some hubs are simply people like ourselves that have purchased a desktop printer like a Makerbot and are offering to print for others to lessen costs of ownership.  Others are people or companies with higher end machines that again are trying to offset costs and employ their machines more efficiently.

I have never sought out a 3d scanning service, but I would expect that some of these hubs that function as businesses might also provide scanning services as part of a 3d model creation environment.  They may also provide help in creating a new 3d model from scanned data.  But my guess is that even if you can't find a printer that also provides scanning, you could get a lead on somebody that does do scanning.

Do some internet searching, and take a look around the Shapeways site or the 3d Hubs site.  If you find a local printer you can often communicate directly with them to discuss your project and their capabilities.  

And I think you should also take a look at the changes you want to make (especially coupler openings) and ask if it isn't something you could as easily accomplish with some styrene sheet and CA at a significantly reduced time and dollar cost.  I absolutely enjoy creating 3d modeling in Autocad, and I love the "perfection" of producing things from that 3d model, but sometimes the process is more complicated than need be.

I hope this helps somewhat, and I'll be happy to answer any other questions as time and knowledge permits.

Jim

One big issue that's preventing me from making things is that people generally aren't willing to share their drawing files or sources for the dimensions. I've used AutoCAD quite a bit and Shapeways will accept files drawn in AutoCAD. The problem is that there's a lack of info on drawings for the prototypes. There's a lot of N and HO scalers on Shapeways with great detail parts or entire shells, but they've all told me that they won't share their files or sources. This is understandable because Shapeways does give some money to users whenever someone buys a product they designed. And, since the files would have to be modified to rescale and modify them to fit O scale counterparts, no one is willing to do it themselves. This really frustrates me because I'm willing to sit and measure my models and change the drawing files to make these custom parts fit, but I have no way of getting accurate dimensions short of walking into a train yard and climbing on an engine to measure it myself.

Now, if there was a way to get blueprints for things like locomotive bodies, or people were more willing to help others instead of going only for money, we might be able to see some pretty incredible stuff coming from websites like Shapeways.

Trevor

It sounds like Autocad converts the raw scan data into a .rcs file that can be interpreted by program.  So I'm guessing that the raw data is about as useful as the .stl files that Naveen encountered.  

Once Autocad converts the raw data, the .rcs (not directly editable) file can be attached to a .dwg (fully editable) file and traced/copied/referenced.  Once the .dwg file is completed, it can then be exported as an .stl file which is what the 3d printer software will interpret.

SolidWorks probably has a similar process although the resulting file converted from raw data may have a different file "type" name.

And I pretty much agree with everything Naveen says about a preference for just starting from scratch based on 2d dimensional data and also data sources. Raster images (think jpg. files) and vector images (think pdf's or other cad files) can be attached or imported into drawings and scaled and traced, or raw dimensions taken off drawings or objects can be entered and manipulated.

Jim

Well you've just identified the number one problem, finding accurate data to start from.  I've been after technical info for EMD htc-1 trucks for years so I could accurately model them in 3d cad.  EMD claims not to have drawings (bs!) and search after search of the internet yields almost nothing.

Scaling an existing 3d model to go from HO to O scale in my opinion is about a 30 second exercise.  It's either done in the cad file before it's exported to .stl, or the scale factor can be set in the printer software when the tool paths are created.  Either way the only real concern is layer or wall thicknesses based on the capabilities of the printer being used.  Going bigger shouldn't present much of a problem.  That's the beauty of 3d printing.  The same model can easily create output in multiple scales without any of the additional tooling required for the same results with injection molding.

I can't really fault people for keeping their 3d data close to the vest.  Information sources aren't quite so proprietary, but if the information is a result of a lot of field research, then that is significant time invested.  The multiscale capability of 3d printing is concurrently a disincentive for open cooperation.  By default you are a competitor because I can make my product for any market.  It's not quite like striking an agreement with Athearn to share r&d data with them producing for the N and HO markets and you producing only for the O market.

All that being said, I feel your pain.  Rebel Locomotives on Shapeways (nee Railflyer) has a lot of great cad work for diesel parts, including my coveted htc-1 sideframes and traction motor housings.  He got the information from somewhere, but I wouldn't expect him to offer it to me for free.

And there actually are plenty of sources of technical data available.  It's just a matter of finding it.  Car and Locomotive Cyclopedias have tons of prototype data in them, including beautiful drawings of the EMD Blomberg sideframe.  Kalmbach publishes diesel and steam cylcopedias that have varying levels of accurate line drawings for locomotives.  And in my case they were available at the local library.  Many manufacturers and railroads have technical data available on their websites.  And print magazines often feature technical drawings, or at least they used to.  Mainline Modeler in particular featured drawings all the time.  Older Model Railroader often had prototype drawings.  RMC did and still does feature a fair amount of prototype drawings.  So the info is frequently out there somewhere, but not always.

Jim

Here is an example of what I did to get 2-D CAD data. I used to have favorite European electric locomotive when I was in HO scale – Siemens ES64F4, in the colors of the leasing company, Mitsui Rail (MRCE), shown below.

When I moved to O-Scale, I greatly missed not having it available. So I collected a lot of photos & scanned raster images of this book.

I also found a pdf brochure of MRCE, saved the front page, shown below, as a JPEG.

 I then imported the image into AutoCAD & over the next several months, mainly on evenings & weekends, I traced lines & curves over every pixel. I used my HO model & the photos to resolve any blurred areas in the raster file & to also make sure that my assumptions weren’t too far from reality.

Since the image was just of 1 side, based on dimensions from the Siemens mobility website, I scaled & mirrored the ends.

I still haven’t completed the center portion yet but I have what I need should I choose to complete it.

These are just my opinion,

Thanks,

Naveen Rajan

I should have said, that is not quite possible.  You do not buy a copy of the file.  Only the 3d printed output generated from the file.  So you can only enlarge it with the methods already discussed.  Either measure from it and create a new 3d cad file, or have it scanned and create a new 3d cad file.  With Shapeways, you never see the 3d cad file behind the printed object.

Jim

I paid over a thousand dollars for my first Sony BluRay player. It took forever to loud some dvd's. What do they cost now? Don

Is there something in particular you are looking for?  You've mentioned pilots.  Can you be more specific.

And while this thread is about 3d printing, in a broader sense it is about how to get things made that aren't readily available from the big manufacturers.  Depending on what you want to build, you might be better off making parts from photo etched metal. In the case of a pilot, rather than modeling the entire pilot area with pilot sheet, steps, and so on, instead create the parts from 2d components.  While still needing to create cad artwork for the etches, it is all 2d instead of 3d.  Much simpler, and the range of softwares that can produce 2d drawings is substantially larger than those with 3d capabilities.  And I guarantee in the end it will be cheaper than creating the 3d cad model and 3d printing it.

Something to think about.

Jim

Jim,

As an example, here are photos of an Atlas 3-rail SW8 switcher pilot, an Atlas 2-rail switcher pilot, and a prototype.  After adding Kadee scale couplers, I would like to replace the 3-rail pilot with something that looks more like the prototype.  As you can see, the 2-rail pilot is only slightly better than the 3-rail version, but still nowhere close to the prototype, so just using the 2-rail pilot would not provide a good result (in particular, the gap between the front steps is just way too wide).

I think in this case it would be simpler, faster, and cheaper to fill that pilot gap with sheet styrene or brass.  Especially since it's an Atlas switcher with the fixed pilot as part of the diecast frame.  

3d printing works wonderfully for building complex shapes that would be difficult to fabricate from individual parts, or for providing short run repetition.  If you were trying to create fixed pilots for typical 3-rail locos where the pilot casting included everything below the frame, then I think 3d printing would make much more sense.  I think for your needs with respect to the Atlas switchers, it's unnecessary overkill.

I'd be more interested in your example with using 3d cad and 3d printing to create the coupler pocket casting.  Presuming I couldn't find one commercially available first.

I don't want to discourage your enthusiasm for 3d printing.  I just don't think it's the most effective solution in this particular instance.  

Jim

From the perspective of my little business, Layout Refinements, I'm very happy to have come across this excellent, edifying conversation. The appearance of the 3-D printer has given me much pause, wondering if they will, in effect, literally, be competitively putting me out of business, considering the amount of hand-crafting that goes into modeling scenes and edifices such as these. These photos are examples of real-life vistas and subjects typical of what I have intend to model for clients. I have been asking myself if a 3-D printer could do it all; if so, why would my skills, effort and time be needed?

Yet, I wonder if the opposite might be true, with such printers allowing that less handmade structures are required, making it possible that any structure that had been photographed, such as those seen below, could be 3-D'd for use by me in scenic vistas and arrays/dioramas? Can you educate me further, please, on what may lie ahead for us in this regard? Would such a printer be capable of reproducing not only structures with the appropriate coloration and weathering but all the foliage and figures and vehicles crafted into any given scene, complete with total, accurate coloration? (Gulp!)

FrankM. (fast becoming a dinosaur?)

(Photo provided by a friend, Jim.)

(Photo provided by a friend, Pete.)

Would each of the elements (sans trains) of this vignette on my layout, for example, be able to be 3-D'd right into existence, all on one base-plate, or at least reproduced piece-by-piece to have at-hand for use?

I don't think you have to worry about becoming obsolete just yet.  As to "what lies ahead" who really knows.  I mean, aren't we supposed to be driving around in flying cars by now?  But the picture is a little clearer as to right now.

Question #1 - No, printers cannot print everything in exact colors and exact weathering as seen in your pictures (they aren't that smart yet).  In fact, color palattes are quite limited.  And the printer neither knows nor cares what color the output should be.  Colors aren't intrinsic to the 3d cad file that the item is printed from.  They're intrinsic to the media used in the printer.  Usually one color at a time, and they don't effectively mix to create the full spectrum of colors.  Items can be printed in multiple colors through various techniques, but effectively only in horizontal layers.  On FDM printers (fused filament types like the Makerbot), the print process can be paused and filament cartridges switched out to other colors.  Printing resumes and anything from that point is printed in the second color.  I'm not sure that SLA type printers (resin cured by laser or UV light) have this capability at all unless the supply of resin can be effectively switched and tightly controlled during the printing process.  I don't think this is feasible for most consumer level printers, if at all.

The numbers on the clock face below are a good example of this technique.  The print was started with white filament, paused, switched to black, and continued.

Question #2 - The success of printing organic objects like trees and foliage will depend on the type of printer being used.  I'd suggest that it's virtually impossible to print something like leaves on a tree, at scale, with an FDM printer.  FDM printers will never handle angles and curves with any degree of refinement.  They do quite well with vertical and horizontal planes and surfaces.  And FDM printers can only print so small because of the mechanics of the printer head that feeds the filament.  Lastly, FDM printed objects can have poor mechanical properties, with strength gained from bonding multiple layers.  A leaf or stem might, at scale, actually be smaller than the thinnest width or layer thickness.  At some point, scaled objects will just be too fragile when printed.

On the other hand, SLA printers can produce some pretty complex objects.  Part of that capability is due to the fact that there is no print head to get in the way.  SLA printers usually produce thinner layers (equaling resolution), and mechanical properties are improved over FDM counterparts.  Can an SLA printer do O scale leaves?  I don't know.  Take a look around the Shapeways site to see examples of what can be printed.

Question #3 - Pretty much everything in your last photo can be 3d printed, even the trains.  Pretty much anything you can imagine (including 1:1 car and jet engine parts and houses) can be 3d printed if you can create a 3d cad model of it and there are no constraints on printer cost, time, and money.  Practically speaking though, some of those things might be better produced using other methods.  Most of us aren't buying $100,000 printers to make trains with.  3d printing is an amazing technology, but it's also slow and expensive.

Most of the consumer level desktop printers that hobbyists might use, like the Makerbot (FDM), Form1 (SLA), or B9 (SLA), have relatively small form factors.  Even the printers I have access to at school only have print volumes approaching 11" x 11" x 11" or so.  The Form1 and B9 print volumes are even smaller.  So you might be able to print that interlocking tower or water tank in whole, but definitely not the coaling tower. And probably none of the above on a desktop SLA printer.

Most if not all of the buildings in your photos would probably be easier, faster, and cheaper to create using other technologies like laser cutting or outright scratchbuilding.  And in many cases the final product will be better as well.  And I have only barely touched on the part about creating the 3d cad model that yields the 3d print.  Either you know the software already (awesome and fortuitous), you learn the software (not bad but time consuming), or you pay somebody to use the software for you (obviously less than ideal).  About the only advantage with 3d printing with current technologies is the ability to create multiple copies of something.  If you can print one, you can print 10.  But is that enough to tip the balance in favor of 3d printing.  Only you can decide that.

If you look at the last paragraph, it pretty much describes the environment we are already in.  At least in the present.  So take solace in the knowledge that you are still relevant.  For now, think of 3d printing as another tool and resource rather than an end all absolute.  My guess is you already employ multiple methods and techniques to produce your scenes.  Use 3d printing as a tool when it is appropriate, and select other options when it is not.

Jim

I don't think management will allow a link ( correct me if i am wrong) but Jay C has posted some truly spectacular parts using 3d printer and lost-plastic casting techniques elsewhere.  Look for an O Scale forum.

The photos he posted show up triple size, and are gorgeous.  He says he used low resolution, and apologized for graininess.  I don't see much graniness. They are the best On3 trucks I have yet seen.