Saturday, October 30, 2021

CN 65ft Gondola series 156000 - S-2 Barber Trucks

For my 3D printed projects, I need S-2 Barber trucks... a lot of them. To the point it would be hard and costly to source them. About 2 years ago, I designed one in HO scale to 3D print and now the project has reached fruition.

Vitruvius, the famous Roman architect how collected in a series of books the best practices of Antiquity's architecture synthetized every design approach to three basic principles: Firmitas, Utilitas and Venustas. In the minds of the Ancients, any object failing to reach a balance between these fundamentals was considered subpar. In modern language, we would translate these principles as Solidity (robustness, durability, etc.), Functionality (usefulness, practicality, ergonomy, etc.) and Beauty (aesthetics, harmony, etc.). Thousands of years later, these evident truths hold true and I challenge you to visit any ancient culture around the globe. You'll find them applied to a myriad of objects and buildings. In my eyes, it is the greatest lesson for any person dabbling into "building stuff". Among all the modern claptrap and buzzwords our modern world love to bath itself in to give us the impression were are so "different" and "freed" from millenia of human history, they are still consciously and unconsciously applied to judge any built (virtual or real) things around us. Isn't it ironic that cultures known for their love of flourishes in their own designs would still be able to streamline a complex thought process to just three poetic words?

So, with the history lesson behind us, how does this apply to a S-2 Barber truck?

Firmitas: these trucks must be sturdy and a limitation of 3D printing is the use of resin which is, by nature, brittle. Not good for a truck that must be pried open to insert wheelsets and which takes a lot of abuse during running sessions. I could only play with two parameters: finding a decent resin that isn't brittle and designing the part to be able to sustain use and abuse. I addressed the last issue by making the sideframes thickness than what you generally see on plastic trucks. I got the idea from trucks manufactured by Rapido. It prompted me to look at prototype drawings and it became quite evident I could make them thick enough to be stronger (and look better). A mistake I made though was to make the bolster fancy and full of voids to save on resin. Unfortunately, it made the truck prone to break at the junction with the sideframe which was an important failure. Finally, I settled on a solid rectangular bolster... keep it simple!

Bottom: original design, Top: new simpler design

Resin was another problem. After experimenting failure with trucks printed by a friend with Elegoo Maroom resin, I was encouraged by Dave Bedard to try my hand at Siraya Tech Fast Grey resin which is far less brittle. I got good results, but after a while, I thought printing black trucks would make painting easier later and if I ever sold the gondola kit, people wouldn't have to paint trucks. The only black resin I could get my hand on was Elegoo, which was brittle... Fortunately, I discovered Siraya Tech offered a clear resin called Tenacious. Having printed parts with this resin, I can assure you can bend it quite a lot without failure. However, I find it isn't the best thing in town when it's time to print high quality details. However, the good thing is that you can mix Tenacious into regular resin to improve them. Most people recommend about adding 20% of Tenacious to a given volume of regular resin. And it works well. Certainly, it doesn't make the Elegoo resin bend that much, but it is far less brittle and print beautifully. I found out it was as strong as my custom mix of Siraya Fast Grey + 20% Tenacious, meaning that it was a well-balanced mix.

Utilitas: The trucks have to be free rolling, easy to print/clean and be NMRA compliant as much as possible. Free rolling was achived by following NMRA standards then scaling parts at 101% to take into account resin shrinkage during curing. It may sound strange, but 3D printing with hobby grade printer isn't magically precise. Discrepancies happen and are normal. You must often tweak perfectly scaled dimensions to take into account shrinking. As for ease of print and cleaning, I had to find the right printing orientation and supports to make it be possible. I won't list everything I did, but let's just say it took several weeks to find something I was comfortable with. Honestly, I'll all good 

Details did matter and had to print well

Venustas: Trucks had to be as close as prototype as possible. It meant, having the cast lettering of the sideframes and perfectly printed roller bearing details. All these things were easy to model in 3D, but required a lot of experiment to make them print neatly. Add supports in the wrong place and kiss goodbye your details... Another major aesthetic point was the resin color. I once painted grey trucks in black and weathered them only to find out later the paint started to rub off with handling, making them unrealistic and ugly. If a truck is already black, if the weathering wear off, it is less noticeable. Also, I have to take into account many modellers aren't into weathering. Offering black trucks is much more versatile and make assembly and painting much straight forward. As you remember, I recently commented on how designing something for others greatly reduce your level of tolerance to unacceptable compromises.

At this point, I think you can get a grasp that dealing with the 3 principles of design can't be done in silo. They all inform each other and each decision is a compromise between two or even three of them. The result is never "perfect", but need to be a harmonious sum of all parts involved, which Ancient Greeks used to call Symmetria... a much more complex idea than the modern "Symmetria" means nowadays (which is basically dumbed down to mirror effect).

Sunday, October 24, 2021

Fences at Reynolds Cable

On the layout, Wieland is a key element where a lot of operations happens. However, space is at premium and most customers and sidings are implied rather than fully modelled. For this reason, the way the short pieces of track connect with the mainline must required a good deal of attention to create some distance and add a layer of plausibility.

The lumber transload is separated from the mainline by a marshy fields, but Reynolds Cable is right by the mainline in an alcove. Early on, we were decided to build a fence like on the prototype to add a visual distinction between elements. However, we took our sweet time getting there. 

During the last few days, Jérôme soldiered on with this project. Before the pandemic, he already created a prototype of movable chain link fence doors using bamboo skewers, copper wire (hinges) and Alkem Scale Model photoetch stainless chain link kits. In my eyes, his prototype was both realistic and strong enough to support manual operation. It was just a matter of installing it on the layout and complete the fence.

The Alkem kits are extremely well designed. Certainly, not as fine as wedding veil, but much more durable, great looking and easier to handle. At some point, a balance must be reached between realism and practicality. Alkem hits that sweet spot its product.

Phosphore-bronze wire was used to add rigidity to the photoeched panels and provide mounting pins. Fences were then painted with Krylon metallic paint and a dash of light grey primer to replicate aged galvanized steel.

When in place, Jérôme added a few Super Trees to mock up the scene. I have no doubt when the photo backdrop, bushes and trees will be in place, this scene will be an attractive spot to switch. Once again, it shows us very little is required if details are done right.

Wednesday, October 20, 2021

CN 65ft Gondola series 156000 - A Proof of Concept

Work is progressing on the CN 65ft gondola. I need them on the layout and the goal is to use them as a first "product" to fine tune my printing skills. The GTR caboose is full of challenges and I'll wait a little bit before doing anything serious with it. On the other hand, this gondola is so simple it's easy to figure out what works and what doesn't.

So far, I'm glad to report printing the ladders in place is indeed feasible. They look good, are sturdy enough to survive normal handling conditions and are rather quite fine in details. The next challenge is understanding why I get bad definitions on the B-end which is printed first. For some reason, details are softer and less defined there. Worst, some resin has a tendency to cure in the rib, creating an unsightly blob. I don't like it and I'm pretty sure nobody would. On a positive note, the decals are currenly printing.

Interestingly enough, my friend, using a different resin, got perfect results three months ago with exactly the same model. It makes me wonder if the custom mix resin I use is the culprit. I'll do some more testing this weekend to find out what's the deal. If I can overcome this difficulty, the kit will be ready to print.

That raises a few legitimate questions. Should I provide a somewhat complete kit including preformed grabirons, wire and metal stirrups or leave it to the modeller?

Under normal circumstances, I would have provided a complete kit, but some parts, like A-Line stirrup steps are getting hard to find in Canada. As we all know, importing stuff from the United States now cost an arm and a leg.

For this reason, I think the gondola should only provide the 1-part  shell, brake wheel, trucks and decals. Leaving all the other small metal parts to the discretion of the builder.

As for assembly, I built three of them in an afternoon for the club layout (yes, I decided to repair the B-end the old fashion way!) and must say they are quick to build. Sanding and preparation is straightforward. In that regard, I made sure the printing supports are easy to remove and can be sanded flush without ruining details. You ten drill 20 holes for 10 standard 18" Tichy grabirons and 8 holes for A-line stirrups. Glue the brake rod made of a bit of wire then add the brake wheel. If you want it, add brake rigging to the underframe, glue Kadee coupler boxes of your choice then add two cut levers.

Adding weight is faster and easier if your work with steel shots. Balancing wheel weights work too but may require trimming. I've personally went the lead wheel weights way and added about 3 oz. to the car.

Then paint, decal, weather and enjoy!

Saturday, October 16, 2021

Fiat Lux

Lighting upgrade was completed on the Clermont section last summer, but it only made it apparent how dark was the other room representing Villeneuve.

New lighting gives layout a museum quality...

Not so long ago, we added under cabinet LED pucks on the ceiling to provide a little bit more lighting. Unfortunately, it proved to be largely unsufficient, particularly when the room ambiant lights are turn off. Several LED products were tried, but most had horrible CRI or were underpowered compared to their official description. At last, we decided to mount a 3500K T8 fluorescent tube like everybody does. This time, illumination was adequate. Certainly, color rendition wasn't pitch perfect, but good enough to fit our needs. We made sure to use the same type of tubes used in the other room.

With a very low ceiling, we had to keep the valence height at a minimum, about 2½'' high. It was made using a white MDF trim moulding anchored on the ceiling with L-shaped metal brackets. With a total cost well under $200, it is the most efficient light rig we have experimented with so far.  We just need a way to figure out how to block light bleeding between the ceiling and the valence.

Many could argue we could do better and I certainly agree. However, given we have at best  about 4 hours of club activities per week, we have to choose carefully the hills on which we to die. Adequate lighting was certainly worth our time and money. Museum quality lighting would probably take months to put together with the risk of being unsastified with the final results. As you can see, photographic evidences show that the new system has drasticall improved the appearance of the layout. It is now framed better and with ambient lights off, it draws the eyes to the layout itself. It also helped to reduce a lot of glare, which was a chief concern. Better, illumination is much better and it shows in pictures.

A slight color shift was observed on CN orange equipment, but not enough to be a big concern given most of our rolling stock is black, grey or brown. Interestingly enough, the same tubes in the other room don't produce this color shift which makes me wonder if the absence of a plastic lense or age of tubes have something to do with it. It could also be pure perception due to how our brain function. When imported in Photoshop, it was quite clear photos under the original lighting were extremely yellow and required a lot of color balance tweaking to look half decent. On the other hand, pictures shot under the new lighting are much closer to reality and didn't even require that I tweak the color balance. And keep in mind it was done only with 3500K fluorescent.

One thing the new lighting makes obvious is the need to improve the backdrop in that room. With more light, all the defects show up in the worst way possible... consider it a good reason to move forward with scenery there in the future!

Original lighting mainly dependent on room light fixtures

New fluorescent lighting

Meanwhile, it's time to go back to Clermont to continue scenery work at Rivière Malbaie. I'm probably optimistic, but I'd like the river and village scene to be done by Christmas, leaving only the ground around Donohue to be done.

Sunday, October 3, 2021

Water for Rivière Malbaie

With club meetings resuming since earlier this month, it is time to address the remaining bare plywood visible in Clermont. We have already started to work on the village scene which morphed back into the initial concept of a farm, however, I felt really bad about the dry Rivière Malbaie riverbed. It has been brown latex for years and its time to move forward.

Rivière Malbaie in 2014

Having the new backdrop in place makes it easier to find the correct colors for water. From what I can see, it's a complex patchwork of browns, dark greys and dirty bluish hues. It's not a gradient, like a lake or a calm river, but more splotchy and impetuous. Indeed, Rivière Malbaie has a strong current during spring.

Rivière Malbaie as it looked from 2017 to 2021

Initially, my idea was to paint the riverbed and poor some Magic Water resin or other clear resin products. However, I feel it could be a real mess and I don't like techniques that require dealing with chemicals and wish the results will be great. I prefer to build up the effect at my pace with more forgiving methods.

Painted rivedbed and drybrushed stones

Once such method is to paint the riverbed then cover it with several coat of gloss varnish and had waves using acrylic gel. It works well and I used it on various projects. However, I fear it would look right for a spring river. More experiments to be done, but more about it later...

The first step was to repaint the riverbed using latex interior paint. The first layer was a generous coat of dark brown all over the plywood to merged together the disparate materials. While the paint was still wet, I stippled and mixed dark gray paint in the middle of the river to create an illusion of depth. The paint was feathered into the brown and blended using a 2" brush. Rotating motions in the current direction created the illusion of moving waters. Once again, I used the backdrop photo to guide my efforts.

The last coat of paint, also wet blended, was a kind of dark turquoise bluish paint. In a few spots, I applied and mixed it to create variation. These spots, on the pictures, are generally near the crest of waves. It gives the impression the sky is reflected on the water.

Then, I looked at the stone embankments. They were glued with PVA glue years ago and it made them look dark. It was unrealistic. Thus, I drybrushed them with a fan brush and a beige color. It blended together the stones by minimizing their extreme color variation. It looked better, but it was not enough. Using cream colored paint, I again drybrushed the stones, adding highlights. Now it did look much better. I also drybrushed the old tree trunk laying under the bridge. It gave it a natural bleached driftwood appearance.

Lesson learned: using natural material such as stone is great, but you have to come back and alter the colors. First, because glue make them appear darker. Second, because interior layout lighting can't compete with sunlight and everything looks darker. Third, you need to blend the natural tones with a coherent and limited palette of colors on the layout.

As for the water, I don't believe gloss varnish will be enough to create an ondulating surface. Also, the plywood texture still shows through the paint. It looks unrealistic  and no coat of varnish will take care of it. It thus means I must fill these cracks and think about a new strategy.

I looked at various water techniques online and I feel the toilet paper soaked in diluted PVA glue should give the results I want. It is easy to apply, you have plenty of time to work with the paper to create realistic waves and no risk to ruin the layout with resin accident. Once it will be dry, I will paint it exactly like I did before because it looked great. Then, several coat of varnish, acrylic gel and drybrushed white paint will take care of the rest.

Friday, October 1, 2021

A New Approach at Extreme(ly) Realistic Weathering


Just a reminder for tomorrow's clinic at Hindsight 20/20. I'll be presenting at 5:45 PM my recent weathering project of an old ex-CN snow plow stored in Clermont, QC.

We will venture into texture and effects rarely used in model railroading but common in the wargaming and armour modelling communities.