Printing a steam locomotive!

Some of you might be interested in these:
http://www.trainboard.com/grapevine...6-Pickles-and-More!-New-stuff-from-CG-N-Scale

 

Totally! One might just need to add temporary supports though (think of a sprue built onto the model). Parts need to be designed not just to survive printing and handling, but also shipping (shipping without the advantage of a custom fitted box).

This is kinda my feel as well - if that doesn't work for people though there is nothing to stop me from offering a "handrail free" version of each model as well. The time required to modify the digital design to remove the handrails would be measured in minutes not hours.

First off, I will readily admit that your technical description shows that you know way more about the hardware than I do. That said, just reading over the datasheet that Projet posts on their website for their HD 3000Plus (the FUD machine), I interpret that they can build in layers down to as thin as 16 microns (0.00063"). Am I reading this wrong?

It certainly can't hurt! Painting has actually saved several models that I thought were looking too rough, I didn't need to dip them or anything.


All that said, I'm not really sure what you guys mean by "stepping;" if you mean "stepping caused by Shapeways as each layer of material is put down," then assuming the surface that you are referring to is not covered by the wax support material (the domes for example) and that you are working with FUD, there is none. In these cases this material can be every bit as smooth as styrene (with no painting or smoothing).

If you mean "stepping" as caused by the fact that the computer software we generally use defines a 'circle' as a 25 sided polygon instead of a true 'circle,' Then the smoothness you can achieve is only limited by the time you are willing to put into making your model and the maximum file size that your computer can handle.

I hope that answers a few questions

 

Nice job on the loco.

It's possible to get smooth curves and compound curves using other software. One of my favorites is called Moment of Inspiration (MoI). It's a NURBS modeler, which means surfaces are built up using NURBS curves, rather than pushing polygons around as Sketchup does. MoI isn't free, but at $300 it's a lot cheaper than most modeling programs. It was created by the same guy who programmed Rhino3d which sells for $995. I got hooked on Rhino back in the late '90s when they released a free beta version of the program. But when the full version came out there was no way I was going to shell out a thousand clams for it.

One drawback to NURBS modeling is that you have to remember to "close" the model so that it's a solid object. 3D printers can't handle surface models that aren't closed.

Steve S

 

For boxcars, you could always design the walls, roof, and bottom separately and lay them out flat. Assuming they print them horizontally, that should reduce any banding. You might also contact other companies to see if they let you specify orientation.

Steve S

 

+1! I'd love to see this. I may have to try my hand at a add on C&0 Elesco setup for the Kato Mike's.

 

There is still some problems with resolution with Shapeways, irregardless of the drawing program you use. When dealing with curves there will still be "stepping" involved. Until Shapeways upgrades their machines to provide for a better resolution the "stepping" will still be a problem, though with better programs to work with it may be possible to make the "stepping" a little less noticeable.

Referring back to a post I saw about a GE E33C being in the works I also noticed that the New Haven EF3 had been mentioned but on the Shapeways site all that is shown is the 3D rendering, not any good shots of an actual shell (still haven't seen a really good shot of the E33C shell that shows close up all the detail). I would like to see just how well Shapeways can render one of these beauties because it would be a good test in regards to whether or not curved parts can be rendered correctly... which may be why only 3D drawings are being shown. I know of some other models in the works and they are being done with high quality software but until the "stepping" is able to be resolved these models can not move forward.

I hope that the higher resolution printing is getting closer... because there are a number of potential models that could then be done, greatly expanding what is available in N-scale. Ultimately I believe that this will be what will greatly benefit all of us, providing for additional modelers to be drawn to our scale.

"I may have to try my hand at a add on C&0 Elesco setup for the Kato Mike's."

GHQ's Southern Ms4 details are still available through Republic Locomotive Works and has most of the detailing needed. Additional parts are also available separately. I will be getting a couple sets since I will be adding these details to some of my steam as I progress with the modeling of them.

 

This is probably the best picture of a complex curve that I have - any more detailed exceeds the resolution of my camera. Keep in mind that the grab iron in the foreground has a diameter of 0.020".





As for the AS616, I'll it to my list.

 

Oh, if you want a body that could gather an easy 20 people, do a CB&Q E5, with a life-like e8 chassis in mind. People have been after one for quite some time. There was someone who had done a limited run, but they sold out a while back and have not been reproduced.

 

That's N scale, right? That's pretty darn good. There really isn't any banding visible.

ETA: D'oh! Just realized this is the N scale forum.

Steve S

 

You can see some banding on the boiler but I think only because it's blown up larger than life size and you're looking for it in the first place, if I saw the finished locomotive on a layout I doubt I'd pick how it was made.

 

That's my feel as well - my point throughout was that the banding was due to my digital modeling only. If you used the software steve s mentions, you could give Shapeways perfectly clean compound curves and there would be no banding whatsoever. As for me - I'd prefer spend my money on more trains though

 

"...there would be no banding whatsoever."

As I have already mentioned, the fellow I know who is working on some things has used the better software to provide for a cleaner image but still the banding or "stepping" is present. Not as pronounced but it is still there. The steamer boiler looks great, no two ways about it, and I would be interested in a couple... hope that soon we will be able to get our hands on some so that we can see where things might go.

 

Hmm well you've got me stumped then, I would have thought the software would have solved the problem. Anyway thanks for your interest

Now wouldn't an L-132 to awesome! If they weren't so enormous and complicated I would probably jump on that suggestion - as it stands though I am myself just a bit intimidated. As for the M-64 though, that would make an excellent fantasy railfan locomotive and it looks like there might be a few northerns in my future anyway... hmm... anything Rio Grande does of course get extra points

 

For those who might be interested in how you model with MoI, I threw this together.

To make the boiler, I drew a circle and extruded it. I did the same to make the cylinder for the dome. To top off the dome I drew a curve and revolved it. To make the fillet between the boiler and the dome, I first went into the top view and made a circle a little larger than the diameter of the dome. Then I projected that circle down onto the surface of the boiler creating a saddle-like curve on the boiler. I drew the small curves connecting the saddle curve to the bottom of the dome cylinder. Then I used the 2-rail Sweep function to create the fillet. I used that same saddle curve to cut out a hole in the boiler. If I didn't do this the entire dome would be made solid by the 3D printer, which increases the cost.



As I mentioned earlier, this surface needs to be closed for a 3D printer to create a body shell of it. Fortunately MoI has an Offset feature that creates a duplicate surface that is offset from the original by a specified distance. This feature works fine for simple shapes, but it had trouble creating the fillet, so I had to model that by hand using the same basic technique that I used to make the original one.



You could skip the hassle of modeling the inner fillet and simplify things by using a cylinder to hollow out the dome. This slightly increases the amount of material needed to make the object, but it's pretty minimal.


Steve S

 

I just realized that I didn't model the fillet properly. The bottom edge of the dome cylinder should also have the saddle shape instead of being straight. To model it I first had to extend the dome cylinder down a bit. Then I made copies of both the boiler and the dome cylinder. I enlarged both of these copies a little bit. I then used the enlarged copy of the boiler to cut the original dome cylinder. And I used the enlarged copy of the dome cylinder to cut the original boiler. I then deleted all the unnecessary stuff and I was left with a uniform gap between the boiler and the dome cylinder.
To make the fillet, I only needed to draw the one small curve for the 2-rail Sweep because the size of the fillet is the same all the way around.



Steve S

 

As for the stack and dome banding/stepping, I think it could easily be taken care of with fine sandpaper and the finer types of Mr. Surfacer, if needed. Otherwise, if they were made separately on a sprue, it would give you the ability to vary stack and dome types. The enterprising small builder (read as randgust, et al) would get the prototype parts cast in Britannia metal and add a plug to them. We need all the weight possible for teakettles. Those plugs would match existing sockets in the shell.

BTW, if you didn't look yet, that's Mark Watson's batmobile.

 

In order to avoid confusion, I think the proper term for the problem that Matt experienced should be "faceting." The term banding would apply to the layered effect that comes from the printing process.

Steve S

 

Matthew, out of curiosity, did you clean the wax off of the FUD parts before painting?

 

Thanks for clarifying that distinction steve - I figured that there had to be a better word out there, I just couldn't think of it


Yea you definitely need to clean FUD in order to get paint to adhere well. I generally run my models under the hottest tap water that I can manage while pausing to scrub them lightly under soapy water. If you plan to order many FUD models then I would suggest purchasing a cheap ($25) ultrasonic jewelry cleaner. I have one; it takes most of the effort out of the process as well as ensuring than parts do not accidentally break.