The billboard feature of the original "Curved Building" model took a bit longer than expected. Part of the problem was selecting the image. I started with Diet Rite because for some reason that is associated with the 70s in my mind. I just couldn't find a good period image. My spouse suggested the "You've Come A Long Way Baby" cigarette ads from the 1970s where Women's Liberation was equated with smoking (Virginia Slims I think). I played with a car ad just to emphasize how amazingly crappy American cars were in the late 70s. Frankly, every product the USA produced in the late 70s was crap.

Reply and let me know if you think of some uSA product that isn't better now and wasn't better before the 70s... What peaked in the 70s other than pollution and municipal bankruptcies?

I also included the 3D render from the model before printing as the last image.

Design Tips For FDM Printing

Fused Deposition Modeling (FDM) is an additive manufacturing technique. The other popular consumer level technology is resin printing. These tips are all about designing models for FDM printing.

Design For Your Printer

I always print with Polylactide (PLA) plastic which is produced from renewable resources and decomposes naturally. It I usually print 0.2mm layer heights with a 0.4mm nozzle.

Real in.	Real mm	HO in.	HO ft.	N in.	N ft.
--	0.1	0.34.	0.03.	0.58	0.05
0.01	0.20	0.69.	0.06	1.17	0.10
0.02	0.40	1.37.	0.11	2.33	0.19
0.03	0.80	2.74.	0.23	4.66	0.39
0.08	2.00	6.85.	0.57	11.65	0.97
0.35	9	30.83.	2.57	52.44	4.37

No matter what software/tools are used to create a 3D model, the model must be "sliced" via slicer software that converts the model into paths for the print head to follow while depositing the plastic for each layer.

(See Image 1 [animated gif] of this post)

Lie To The Slicer

I tell the slicer that the nozzle is 0.39mm wide and layer heights are 0.2mm. I use 0.39mm width so that lines exactly 0.4mm wide will not be discarded by the slicer software. When the slicer software treats 0.4mm as the minimum, many lines don't print presumably due to rounding errors (precision limits) in the slicer.

Do Not Model Anything Horizontally Thinner Than Nozzle

Any model geometry that is thinner on the X or Y axis than the nozzle will either not print, print too wide, or only partially print.

Model Horizontal Detail in Multiples of Nozzle Width

When using a 0.4mm nozzle, lines/walls that are 0.4mm thick are great. 0.8mm thick lines/walls are great. There is no 0.3mm wide. There is no 0.6mm wide. There is no 0.9mm wide. Design for 0.4mm, 0.8mm, or 1.2mm. WHen larger than 1.2mm, the slicer software will generate an "infill pattern" comprised of adjacent parallel diagonal lines, and the print result will usually turn out fine. You may find the printer trying to fill tiny gaps in the infill pattern, and that slows the print but seldom causes quality problems.

Do Not Model Anything Vertically Thinner Than Layer Height

When printing 0.2mm layers, anything thinner than 0.2mm on the Z (vertical) axis will either not print or only partially print.

Model Vertical Detail in Multiples Of Layer Height

Details 0.2mm high are great. Details 0.4mm high are great. There is no 01.mm, 0.3mm, or 0.5mm!

Avoid long "Lofts" And Non-straight Lofts

"Lofts" or "bridges" are parts of your model where the printer needs to extrude plastic over air. Short gaps that can be bridged with straight lines are usually fine. I've successfully printed as far as 20mm over air under the right circumstances.

(See image 2)

It's common to print lofts over window openings. Lofts that require non-straight lines of excessive cantilever over air are unlikely to print well.

(See image 3)

The solution is to print parts separately. For this model, three parts are enough to avoid lofts with the exception of short spans over windows. All large horizontal surfaces print directly on the print plate.

Extremely Small Relief Detail Is Possible

Features the stick out or stick up need to be multiples of teh nozzle width and layer height. However, space between features that stick out can be as thin as you want. Consider bricks: each brick can be 0.8mm by 1.8mm and raised 0.2mm - the gaps between bricks, the mortar, can be 0.01mm wide if you want (roughly half an inch in N Scale).

(See image 4)

---- CRITICAL ----
Avoid Non-manifold Geometry. Imagine filling the volume of your model with water. If the water could get everywhere in your model without leaks, the model is manifold.
Most slicer software will malfunction when slicing non-manifold models.

Avoid Online STL Files

What!? That's right. Trying to use or reuse online STLs is a path to frustration and low quality. Why?

• STL files are difficult to edit. The format is designed for slicers and not humans.
• The model used to produce the STL files almost never adheres to the guidance above. If the model has thin walls or walls that are not multiples of your print nozzle's width, it will be more trouble to correct the model than to start from scratch. If the model is non-manifold, there may be no practical way to correct the problem.
• In places where separately printed parts join, you must leave a gap in the model so the printed parts have room for glue and account for slight variations in printed line width. I usually design a 0.2mm gap between parts that need to join.
• The STL file may have never been printed on any printer using any technology. Its problems are unknown.
• The model may be the "wrong" scale. As soon as you try to resize the model, the model is very unlikely to adhere to the guidance above.
  • Shrinking a model produces elements too thin or too flimsy to print.
  • Shrinking the model makes designed in gaps and tolerances too small. The resulting parts no longer fit against each other, or space glue forces the parts apart.
  • Shrinking a model may result in connections between parts that have insufficient surface area for glue or are too fragile.
  • Enlarging makes small details like window mullions and railings too thick.
  • Enlarging make a loft that your printer could have handled into a bridge too far.
  • Enlarging makes gaps and tolerances that were perfect for snug fit into unsightly and structurally weak joins.
  • Trying to fix problems after resizing is usually far more work than starting from scratch.

Design For Painting And Assembly

I design models so that parts that will be painted different colors are printed separately. This enables me to put all the parts that are the same color into my makeshift paint boot and apply rattle can spray paint. Then I swap with the parts and spray a different color.

The separate parts are designed to interconnect in ways that simplify assembly.

Assure Parts Only Align And Fit One Way

Key the parts so they interlock. Design parts that join to conceal seams and prevent sliding out of place while glue dries. The parts should "snap" into place together and resist being separated again.

(See image 5 and 6)

Most consumer FDM printers these days can print with tolerances of 0.05mm or smaller. WIth that little separation between parts, I sometimes avoid using glue. I couldn't get the interlocking parts to separate again without breaking them.

---- BONUS ----
I'm always frustrated when an assembled kit has openings at corners where I didn't glue them tightly, or I glued them slightly out of square.
I hate it when parts slide along joints to become imperfectly aligned while glue dries.
Your own designs can avoid those assembly pitfalls. Assembly becomes a foolproof endeavor When parts only fit one way, cannot slide against each other, have ample surface area for glue, and interlock to form perfect angles.

As promised, here is the printed, painted, and assembled model of "Sunshine Cab Company".

The pictures show the SketchUp model and SketchUp's AI render of the model. The parts have all been printed and painted. I weathered and sprayed dull-coat ten minutes ago. As soon as my clear polystyrene window glazing arrives from Amazon, I'll add window glazing and assemble the structure. I'll post images of the final result.

Since my last post I learned that track doesn’t have to be expensive and picked up flex track the next day. But I still want to pursue my dead track idea.

I am new to the train hobby but come from diy and electronics. So I’m making the control logic myself. I will be posting this. I will have nfc/rfid tags under the track so I can have the train recognize what section it is in. This gets sent back to the controller. When a train is waiting at a station another controller turns on the power. Allowing the train to charge. But in my control logic I will convert this to fuel. So they fuel up and use this virtual fuel during operation.

Is this the best way to do it or the cheapest? no, but the track is cheap. I got about 1 meter of flex track for €7-8 while the same length in printed track costs only €1-2.

I’m happy with the current result but still working on it. Want to also make some tools to make it easier to lay out and glue. It can also plug into the original pico track like my last pieces. Ballast also works for the flattrack.

I will be publishing the files once I’m happy with them. Need to fix some things. Plus if you use fusion 360, it’s parameterized, so you just enter the length in amount of sleepers and it generates that.

That was my yap for now

enjoy the rest of your moment

After making a whole new level with 3D printed viaducts I realized that despite the Kato 381mm radius my large trains still rub, including my big UP DD40X on the inside.

I decided to draw some new custom viaducts in CAD and I'm just beginning the process of replacing these all to allow for more clearance for potential Japanese trains or a big boy. These will look more like a girder.

The difference in clearance is huge. I think they look okay and will just need some weathering to spruce it up.

Pic 2 shows the new one on the left, next to the old oversize viaduct on the right. Pic 3 is the new clearance!

No rubbing now!

Here are three printed "New Haven Style Catenary Poles". One is 4mm webbing (2 scale ft), one is 3mm webbing (1.5 scale ft), and the last is 2mm (1 scale foot).

Keep in mind that on your computer screen, these images are likely 10 times larger (or more) than they are on a layout. At a normal viewing distance on the layout, the 3mm webbing is quite convincing IMHO. Weathering and touch-up painting can make the webbing pop even more.

Here are the STLs. There are more images at the same site.

I just bought my first model train today. It’s a Piko RH2016 Hercules, but analog. H0 scale. I honestly was just the cheapest set and I’m completely new to this, but not to printing and electronics.

I find track to be really expensive but got some in the starter set. I’ll connect this system to an ESP32 and electronics to have my own control system/logic and and battery power, charging from the rail.

I 3D printed the rails to work with the piko system I got with the train. I expected it to work but couldn’t have wished for a better fit. Still waiting on other filament colors to come in.

The model is completely parameterized so I can just change the sleeper count to change the length. Yet to make turns, which is nice cause I have complete freedom.

I want to use powered metal rail at the stations and will convert battery level to fuel. So they fuel up at the stations. Use the charge controller light signal to enable engine noise in the future maybe?

That was it for now, wish me luck!

The station module for my elevated railway project has finally made it through test printing.

And, rather suspiciously, most of it actually worked after the first print. :)

I assembled the full section to check the proportions, platform clearance, canopy structure, railings and how the station building sits against the track. I also put one of the IRT cars and the OO mate on it because CAD has a remarkable ability to convince you that absolutely everything is the correct size.

The IRT car is still firmly in the "resting on the platform looking useful" stage, as I’m still hunting for period-appropriate bogies, preferably motorized ones, that will actually fit it.

If anyone knows of a suitable HO powered bogie or donor chassis for an early IRT car, I’m very much open to suggestions.

The whole elevated railway system was designed specifically around FDM printing, so all of the parts print without supports. That includes the building sections, roof, canopy structure and all of the smaller architectural details.

There are still a couple of small things I want to clean up before I call the files finished, but nothing in the test print requires a major redesign.

The elevated railway itself is modular, so the station can be inserted between normal track sections rather than being a separate fixed layout piece.

Next I’m planning a proper station entrance/building at street level and a junction/signal box for the system.

As always, feedback and suggestions are very welcome.

Users bcentsale and 382Whistles expressed interest in New Have Railroad style catenary towers in the comments for https://old.reddit.com/r/nscalemodeltrains/comments/1vr4qk6/a_detail_i_am_proud_of_that_i_forgot_to_post/. I was a bit frustrated with my latest structure model's progress, so I decided to model the New Have style catenary towers in N Scale. The STL file is available.

Doing a bit of an experiment to see if I can 3d print a layout, the first 3 boards are now wired trains are running (btw ignore the missing buffer and coal on the j50, I brought it like that and im going to fix it one day)

This album is progress on my N-Track module. I hope to show the module at a Nov. train show.

• 48"x30"
• Lauan Plywood construction (everything cut from a single 4'x8' sheet. - Soft pine dimensional lumber is used for C-clamp placement and as a base for screwing/mounting the legs.
• All structures, bridges, abutments, railings, fences, etc. are original 3D models printed on a Bambu A1 with PLA 0.4mm nozzle and 0.2mm layer heights. Apart from the Kato track and raw materials, there are no commercial kits or components.
• Rocks are hand carved plaster over foam board insulation.
• The three required community tracks run through the module below grade. The scenery including rocks and structures lifts out easily to access the community tracks.
• The rocky side is the front. The module is setup to be an annex to my regular layout as well as a display unit. The rocks align with scenery on my main layout.
• The track plan is inspired from David K. Smith's Jersey City Industrial

A little update on the 1916 IRT car I’ve been working on for the elevated railway project.

The model is mostly finished now and this is the first time I’ve been able to put the complete body on the elevated structure and station. Seeing it in place definitely helped me catch a few things I still want to tweak.

This test was printed with a 0.4 mm nozzle. I’m actually pretty happy with the rivet and window detail, but I think this model is right on the edge of what makes sense with a 0.4. I’m going to pick up a 0.2 mm nozzle and see if i can print the smaller details, particularly around the windows and front/rear panels.

I also modelled a basic interior since the windows are large enough that an empty shell looked a bit strange.
I’ve decided not to 3D print the running gear. I’ll use ready-made bogies and metal wheelsets instead. At this scale I’d rather have reliable running gear than force everything to be printed.
Still a few adjustments to make, but the car itself is getting pretty close.

I printed the "Corner" building in 2022 on my old Bibo printer. I usually print walls flat on the print bed and then paint and assemble. To achieve a curved wall, I printed the whole facade as one piece with "lofts" over windows and doors.