Wednesday, July 29, 2020

CN Woodchip Car - A Proof of Concept

Yesterday, a few test prints were made to confirm if printing ladders and brake platform directly on the model was both feasible and satisfactory. A big thanks to Guillaume Pedneault for providing these samples with helping me with the final design.

Car ends with ladders and brake apparatus (credit: Guillaume Pedneault)

Be aware the sample weren't cleaned out and got only a quick coat of red primer to make details pop up. I was quite happy to see most details printed well, including the brake wheel and fulcrum. However, we have serious doubt the small chain would survive under normal circumstances when cleaning the model. This will have to be addressed with some kind of support under the chain itself. I still have to figure how it will be done to stay unnoticeable. I could have omitted this detail altogether as I did on my original pre-production model, but it definitely had a level of realism that should be appreciated by modellers.

As for the trucks, I've yet to figure out a few things about NMRA standards and recommended practices. Their information on journal dimensions seems to be a little bit weird and I tried comparing them to Accurail, Rapido and WalthersProto trucks... none of them matched the standard.

Barber S-2 truck sideframe (credit: Guillaume Pedneault)

That said, the printing went well. I can still see a few areas that could be greatly improved on the trucks. However, the person helping me with the prints gave a few good hints and it paid off. The springs looks good even if the sideframes aren't completely cleaned. The cast on lettering is always surprisingly good. Barely readable, but still there. The big challenge is making these trucks run freely.

Meanwhile, I working on a completely new 3D model of 879000-series woodchip car. The original 3D file was full of artifacts due to me not knowing what I was doing. Fortunately, with a better grasp of my tools, it's far easier. Ditching the millimetric template in SketchUp and learning a few new commands is making the work faster and better.

Tuesday, July 28, 2020

CN Woodchip Car - Barber S-2 70-Ton Trucks

I've been trying to source parts for my various modern freight car projects, however, these are sometimes hard to find when you look for nice ones or simply hard to source when you want to buy them in bulk. Given the actual precarious conditions and given it would be foolish to count on unreliable sources, I've decided to design my own version of Barber S-2 70-Ton trucks equipped with roller bearings.

Working with photographs, drawings, NMRA standards and recommended practices and a few HO trucks from various manufacturers, it took me about half a day to come up with something both pleasing and, I hope, will be mechanically sound. I wouldn't have never believed there was so much variation with the S-2 truck family. Honestly, it seems none are born identical. For this reason, I tried to reach once again an acceptable compromise by matching my work as close as possible to what what photographed by Justin Babcock on the prototype a few years ago. Bear in mind it is highly possible this wasn't the freight car's original truck or that great discrepancies can be found among the fleet due to 40 years of maintenance and repair. However, what I design should be quite right at capturing the feel of the real thing.

The model will probably require a few adjustments after conducting rolling tests. For this reason, two designs will be tested. The first is made of two sideframes and a bolster that are glued together. The second one is a 1-part complete truck.

Monday, July 27, 2020

CN Woodchip Car - Third Pre-Production Model

The last two weeks have been spend redrawing completely the 878000-series woodchip car to get it right. It was also the perfect occasion to think about how it would be printed, the level of details and how the thing can be assembled by modellers with various skill levels ranging from newcomers to skilled craftsmen.

This new pre-production model will feature 3 main parts: a 1-part body with sides and underframe details and 2 end parts including ladders and brake platform. Gluing them together will provide a complete model.

The modeller will have to add a few grabirons, glue the brake wheel, cement a brake rod and install standard Kadee couple boxes and S-2 Barber trucks. Ladders, brake housing, brake plaform and fullcrum are already printing and don't require assembly. Same thing for underframe brake reservoir, valve and other such things. I've also made the stirrups molded directly on the shell. I'll see if they are sturdy enough to be worth it, but from what I've seen, they should be fine. That said, more detail driven modellers will have the option to remove the stirrups and replace them with metal ones if desired.

As for brake rigging, I've provided none. The reason is simple: most modellers don't care about that while the ones that do care generally prefer to add the details themselves according to their own personal standard. So instead of providing a half-baked solution to both type of modellers, I prefer to offer a decent underframe upon which one can do exactly what he wants. I'll probably provide a brake rigging diagram for guidance. Also, holes in the underframe are provided to install air lines.  

Weight was certainly a problem for these open cars. While a neat underframe is always cool, these cars can be often seen travelling empty on layouts. Thus, I feel the interior should be as accurate as can be. It means I've design a recessed floor in the underframe which can hold a steel weight that brings the car within NMRA standards. The weight is then covered with a 1mm styrene sheet to hide everything and provide a smooth floor like the prototype.

For couplers, I originally designed quite complicated and nice boxes, but the thing is 3D printed resin can be brittle when you drive a screw in very thin-walled parts. Most of them broke on my pair of preivous pre-production models, so I've been advised to simply swith to Kadee. To be honest, the difference is barely noticeable, even under closer inspection.

So now I only need to get the printed version of these new pre-production models. As you can see, I've had to make a few compromises to take advantage of technology instead of fighting against it. In some sense, it has made the model easier to assemble and sturdier, which is a good thing. I know a few people will want many cars in their fleet and assembly should be relatively straightforward. I'm also happy the model can provide a good starting base for a fully detailed model. I took care to have coupler drawbar supports were required for this reason.

Wednesday, July 22, 2020

CN Woodchip Car - Start From Scratch

I've often wrote about the inability of SketchUp to manage sub-millimetre dimensions. The software is now showing its age, having not been designed to take into account 3D printing back in the early 2000s.

The thing is that as you add parts to your model and merge together shape, small discrepancies creeps here and there. At first, it is unnoticeable. If you enter 0.500mm then you really get 0.500mm, but after a few merges of various components, that becomes 0.498 and some weird angles start to appear. What used to be 90 degrees, will misalign itself on other component. It has been noted by users on SketchUp forums the software have a hard time dealing with very large and very small dimensions.

I've redrawn the model 3 times in the last few days, often from scratch, but always ended up with the same discrepancies. I've worked with CAD softwares for decades now and I now I'm a precision-oriented kind of maniac in this regard. I can't stand odd numbers when dealing with plans and I know, after controlling several parameters, I wasn't the main source of discrepancies.

Most people would think it isn't relevant because most of these errors are several orders of magnitude smaller than what a 3D printer resolution could manage. That's true. However, it becomes a real problem when you try to merge elements together. SketchUp works with faces instead of volumes, so if "parallel" faces aren't really parallel, you can no longer deal with coplanar geometry. In my case, it made it almost impossible to merge together the underframe and walls. This is a big problem because drawing the model using independent components then merging them together later as you go along allows you to manage parts easier instead of fighting with a single part.

Main shell - Almost completed

My first reaction was simply to model the prototype by making it 100 time bigger. That way, I wouldn't have to deal with sub-millimetre dimensions. yet the same issues. Thus, I tried this morning to draw the model in meters. Each millimetre became a meter. In fact, 1.5mm is now 150m. By setting the drawing template to meter eliminates SketchUp inadequacy to address small dimensions. While the idea to enlarge the model was a good one, it only worked when I switched to the metric template (of feet if you draw in imperial units). Thus, the problem isn't about absolute numbers, but related to unit. If you draw 15000mm, you'll get discrepancies. But if you draw 15m, you get absolutely none. Lesson learned! Hope it will help peoples out there using SketchUp! And yes, I know several new intuitive 3D modelling softwares are out there. I'm just trying to make the best of what I have.

Main shell showing the floor depression hiding the metal weight.

That said, redoing the model several times was a good way to improve the design and I'm glad I did. I've reach the point I can redraw almost from memory. Many things have been simplified and I found a way to hide the weight in a way that will look nice when the car is empty while not scrapping all the underframe details. As much as I can, I feel this car must be as straightforward to assemble with being easy to superdetail for modellers wishing it. Maybe I'm wrong to try catering to two different crowds, but this is something worth trying to achieve.

In the next blog about this car, I'll explain how the model will be assembled using - I hope - cool graphics to demonstrate the design.

Friday, July 17, 2020

CN Woodchip Car - Toward Production

Last time I wrote about the CN woodchip car project, it was back in early February when the pre-production models were unveiled during an informal RPM meet in Ontario kindly organized by Hunter Hughson. Since then, COVID-19 happened and my professional workload skyrocketed due to a new exciting heritage project. However, some readers have contacted me about the cars, asking when they would be available. I feel I have a moral obligation to stand by my words and thus decided to make a last push to see this project reach a satisfying conclusion.

However, the  delay isn't a bad thing because it helped me to get access to better technology while addressing issues raised by the pre-production model. With the help of some people in Quebec City area with good 3D printing knowledge it will make developing the project both faster and better. Also, doing thing locally helps me to keep an eye on quality and reduce costs to some extend.

That said, let's address the issues and flaws found in the pre-production models:

First, the underframe small I-beams were extremely brittle. I tried to make them as close to scale as I could and it was a mistake. A few people with knowledge in 3D printing told me to make them rectangular to make them sturdier and easier to print (no reliance on too many supports). I know it is a trade off, but it's the underframe and I know most people won't care as much as a few ones about that. If I was producing a resin craftsman kit, be assured no shortcut would be taken, but it's not the case. I'll live with this decision, knowing it makes the product easier to sell.

Second, I know have access to a larger 3D printer that can print a full car as a single part body. This is a great news because it removes a lot of assembly steps that were a little bit tricky. Assembling an open car from flat panels may be not that great for many people. Also, it makes the car sturdier once again. With that said, I'm experimenting with a shell on which ladders, platforms, stirrups and other details like brake chain and fulcrum are molded on. If it works, I'll go that route, if it don't, nothing loss and I revert to my original option with separate parts.

I've got to be pragmatic because people don't want to assemble models and most people interested in operating these woodchip cars aren't the craftsman crowd. That said, I've taken great care so most details visible are accurate, making it possible for modellers with higher standards to improve by adding additional separate details like brake rigging under the car.

I should have a second pre-production model to show soon. And honestly, I'd like to have everything ready for production starting this September.

I'm also working on CN 156000 series 65ft gondolas. The 3D model is ready for pre-production and decals are made. However, that will happen after the woodchip cars are done once for all.

Monday, July 13, 2020

Repainting Ties

Long time ago, I use to weather track following a basic recipe used by Mike Confalone. It was all about spray painting the ties and rails with a dark brown, generally Krylon camouflage brown, then handbrushing various shades of thinned down beige, brown, black, and grey over the ties. It was supposed to create realistic variations in tone. If your familiar with Mike Confalone's work, you know for sure this technique does work and yield great results... except it never worked for me. Ties ended up covered with blotches, not exactly a visually appealing effect.

Ballasted track at Donohue... not impressive.

It's why most of the layout follow another technique often described on this blog with consist of painting the ties white and applying several oil paint washes to build up the color. I've always enjoyed consistent results with this method and it became a standard.

Unrealistic paint blotches.

However, out of laziness, and because the photobackdrop was already in place at Donohue, I never dared to upgrade the track paint. A few months ago, I simply ballasted a few feet of that track and found the result beyond appalling. It was clear we couldn't go that way.

Tracks in the repainting process

To save ourselves some trouble with spray paint cans, we thought about painting by hand each ties with white paint. It is not the fastest procedure out there, but in a single hour, with two people, we painted about 6 feet of track. Given it was a test, I suspect we will be faster next time and think it will probably take about 2-3 hours to complete the scene. It is certainly painstaking, but it only shows that shortcuts taken a few years weren't exactly the best idea.

Sunday, July 12, 2020

Scenery Progress at Donohue

This year goal seems to complete a sensible amount of scenery in Clermont and particularly Donohue which has been Plywood Central for many, many years... in fact, since May 2014.

Work continue on Rivière Malbaie with some spackling applied to get rid of gaps in the plywood. Once sanded down, the river bed will be painted in various shades to reflect the backdrop photo. I've yet to decided how the water will be done, but it won't be more than 1/8" thick, meaning most effects will be done using various varnishes and clear acrylic gel.

We also painted and installed the illustration board road and parking lot. Once again, the carboard was sprayed with various shaded of automotive gray primers and Krylon camouflage beige. This time, more lighter shades were used to represent an aged pavement. Cracks were added with a black pencil, following prototype pictures. Later, when scenery will be almost done, various powders, including weathering ones and real stone dust will be brushed over the cardboard pavements to add another layer of realism and blend everything together.

I find it easier to glue these roads with acrylic caulk. I'm not a fan of PL300 glues, these being too much overkill for modelling purpose. Once the roads are set in place correctly, I run a bead of carpenter glue along the edges to seal them from future scenery glue and to make sure they won't warp at the perimeter.

Sunday, July 5, 2020

As I continue revisiting several layout concepts so they can fit within my display cabinets, a new vision of Monk Subdivision start to emerge. Simplified and narrower (about 16"), this different take on the project tries to wrap around the corner the entire station, yard and depot. You'll also remark the main line isn't parallel to the walls, but flows diagonally to create space for various railway function (the shops, the depot, etc.)

One big change is to only consider the cabinet area as the main operation focus, the rest of the smaller shelves acting as a simple runaround track. On that loop, one two-ended siding could be added to stage a train or two in advance or act simply as a originating and terminating point for trains.

I also decided to keep the yard as small as I could, simply keeping a long passing track and a smaller two-ended team track serving various customers, including an oil dealer. While studying a few pictures from Monk, it became quite clear a freight house used to stand left from the large depot. It seems to have been a regular standard NTR model and the team track there could handle a few freight cars, maybe 3.

The turntable has also been relocated in a corner so the service tracks could be longer and more realistic. Having the locomotive shops on the opposite side of the layout creates a visual tension on the layout while providing an interesting focal point when entering the room. For the sake of convenience, a small MoW siding was added so some interesting rolling stock such as work trains or cabooses could be displayed there.

The continuous run could be scenicked of left as is. It could also be possible to imagine a small customer such as a feed mill located on the curve for the sake of performing smaller local train operations. But it wouldn't be required per se.

This version of Monk is probably much toned down compared to the original concept which was sound yet a little bit crowded to my taste. The goal was to create the impression of a track surrounded by nature and wilderness as was the NTR. Structures are clustered together, track density kept as low as possible and most secondary trackage would be in bad shape or covered in grass as was the case on the prototype to emphasize the linear nature of the mainline. I once tries with with Villeneuve on the club layout and found it was a neat trick to create the required hierarchy of track use.

I've also experimented with a slightly different version with a completely scenicked continuous run. While it could be interesting, I think this plan kind of thin down the intial focus on the cabinet area.

Friday, July 3, 2020

Avenue Industrielle - Richer Than Previously Thought

Recently, Jérôme sent me a scan of an old 1965 map of Limoilou Yard in high definition. First time I ever saw a good scale drawing of the area before major overhaul occurred in the 1970s. It helped to better understand how the trackage changed near Glassine Canada which, it turns out, was located on what used to be QRL&PCo Limoilou Shops wye. You learn something new everyday.

However, the biggest surprise was the industrial spur along Avenue Industrielle which I have covered on other blog posts in the past. This time, it was really an eye opener. Until now, I only knew about this interesting small industrial park through very unreliable insurance maps and grainy aerial pictures.

While I kind of guessed most elements right, it seems I my original interpretation was right: their was a team track near the roundhouse. For years, I thought it was simply my imagination, but it was indeed real and had a loading ramp.

From a layout perspective, this is quite interesting because it means a diversified traffic occurred there on a daily basis. Thus, a very simple track plan is no excuse to call it boring.

Colorized Limoilou Yard map (credit: Ville de Québec, circa 1965)

But on a surprising note, it seems a big chunk of marshes still existed west to the spur. Indeed, Limoilou yard is built on an ancient swamp called La Canardière where ducks gathered (they still do to some extent). But I wouldn't have guessed parts of that marsh was still surviving as late as the mid-1965.

This little geography fact has a tremendous impact on how I would approach building such a layout. Instead of starting with a flat plank, I'd rather build the entire trackage on an embankment then fill the areas with industrial activities and leaving the natural areas lower. The spur itself would be built upon a dirt and rock embankment full of bushes and weeds, making it clear it was built on previously natural lands.

Revised track plan

Given we always approach small industrial layout from a very dry and rugged perspective, it seems quite refreshing to see it as a small chunk of human activities slow encroaching a natural habitats. It makes for a more dramatic and contrasted visual narration while providing much more historical context and a means to tie together a very long scene (about 20 linear feet).