My passenger car is finished and packed ready to go down to Sacramento to show in the NMRA national show next week. To support my display in Sacramento (and likely in Burnaby this fall, and who knows where else), I threw together a quick 20-page Blurb book on the project. It is laid out like an NMRA merit judging form, and I’ve pasted the bulk of the content below:

Introduction

Surely I could have chosen something simpler for a first passenger car project. Surely I could have found a car where we have actual drawings if not clear photographs. Surely nobody would mind, even, if I bought a kit for a passenger car, lettered it for the Canada Atlantic, and challenged anyone to call me a liar.
I would mind.
The prototype modeling path is a personal choice, and we all must decide how far down the path we will go. For myself, there are so few resources showing the Canada Atlantic as it really existed, that I feel I must replicate ever one of them as closely as I can.
My current modeling project is the town of Pembroke, Ontario, and we have four photographs of the same passenger car on the line. This, then, is the car I need. Substituting another car would be akin to ignoring one of my sources in prototype modeling, and ultimately diminish the total effect.
So, without drawings, photographs, incontrovertible dimensions, without even a car number, I set off to replicate the car and the train, pulling into Pembroke sometime around 1905.

Prototype

My model of Canada Atlantic #2 is based entirely on four photographs that show a combine on the Pembroke Southern. From the photos, it is impossible to discern the number; however an undated roster from the Grand Trunk Railway shows two 1st/2nd/baggage cars that were inherited from the CA — numbers 2 and 4. Assuming our Pembroke Southern car was leased from the CA, which was the operating railroad, and assuming it survived until the GTR roster was compiled, the PS car was either CA 2 or 4. We have a clear photo of CA 4, and it does not match the PS car. So, if all our conjectures are true, then the PS car was CA 2.
Having said that, by the time of the roster, the GTR car was clearly different from the PS car. Notably, on the GTR car, the windows were double, not single. This seems a major structural difference, but all the other dimensions match. It is possible that the car was modified by the GTR or that the roster was incorrect. In any case, this model represents the Pembroke Southern car, which was certainly leased as the PS had no passenger equipment of their own, and the CA seems a likely lessor, and CA #2 is a likely identity.
The length of the PS car was drawn from the photo of the train at Golden Lake and the known length of the bridge spans at Golden Lake. Dimensions such as height and width were standard on the CA to within three inches, and so, I used those standard dimensions where they were applicable.
Where I had no data for the PS car or for the CA, I resorted to industry information. The CA built many of their own coaches, but they also bought cars from Crossen in their earlier years, and Pullman later on. We are fortunate that Ted Rafuse has written an excellent book on the Crossen Car Company, and this served to furnish some typical details.
Further information on typical car construction was gleaned from the Voss book on Railway Car Construction, especially the chapter on a NYC&HRR day coach. The cross-section of the clerestory was especially helpful in getting the roof correct, as was the body bolster drawing. Many other details were pulled from the White’s book on passenger cars and from a visual dictionary for car builders.
Finally the question of colour took months to resolve. Obviously, we have no colour photographs of CA rolling stock. There is one colorized post card that may show some of passenger cars, although the cars are on the Canadian Pacific’s track, and so they are just as likely the CP’s. We also have a newspaper account suggesting that CA passenger trains were “Turkish Rouge.” This note lead me on a lengthy mission that ultimately culminated in replicating an antique cosmetics recipe using the original ingredients; the key ingredient is Alkanet, and it has a beautiful pinky-red colour under some light. I’ve never seen a colour change so dramatically from one light source to the next, however, and the colour of the model depends heavily on the light under which it is viewed.

Construction

Roof and Body

The hard part of passenger car modeling is the roof, specifically the end where it is all compound curves, and not a straight line in sight. As I pondered how to create these ends, I became aware that 3D printing was becoming increasingly capable and within reach. So, I resolved to print the roof.
A common failing of passenger cars, especially those with separate roofs is that the ends of the letterboard are very fragile and never properly engage with the ends of the roof. Printing the letterboards along with the roof meant that the colour separation would have to be masked, but the parts would mate perfectly. Once the decision to mask the letterboards was sealed, it seemed obvious that I may as well print the whole body.
So that is what I did. Working from dimensions pulled from photographs and prototype practice, I developed a 3D computer model of the body and roof. The resulting model contains some 31 thousand vertices, and includes details such as door handles and flag holders. This model was then exported and printed using stereolithography. The stereolithograph required cleaning and sanding to remove evidence of the production process, especially on the roof.
Sadly, the stereolithography process does not facilitate very thin sections, and so, the window sashes could not be printed. However, because I had a computer model, it was a simple matter to create a pattern for a laser cutter, and cut these out by machine. The grooves where the real sashes ride were included in the model, and the laser-cut sashes ride in these grooves as per the prototype (they are fixed in place, however).
For a while I thought I would leave one of the baggage compartment doors open on the model, and to support the thin cross-section, I laser cut these at the same time as the window sashes. As long as I was laser cutting, I also cut all the glazing as well.
The roof was dressed with a number of scratchbuilt detail parts such as the baker heater expansion tanks and stacks, the lamp jacks and the toilet vents. I created these separately rather than integrally with the model either to facilitate colour separations or so that they wouldn’t get in the way of finishing the roof.
The only details left off the body in the computer model were the handrails. The flattened ends of these were included, but the handrails themselves were too fine to print, and so, they were built up in brass.

Underframe and Platforms

The core of the underframe is a 1/8” thick piece of steel, cut to fit precisely into the body. This provides almost the correct amount of weight as well as super-strength to what could be a troublesome floppy piece of modeling if done in wood or styrene.
The steel underframe is covered with a wooden sound-deadening ceiling that I built board-by-board. The needle beams and body bolsters were incorporated into the ceiling as, being frame members, the ceiling would have been built around them.
The platforms, steps and end beams were stereolithographed together with the interior, ensuring that they are straight. They are somewhat vulnerable out there, however, and so, my next car will incorporate more of the steel in the platforms.
Various details, such as the brakes, levers, train line and so on were glued straight to the sound deadening ceiling or between the frame members under the platforms. These details are all made from commercial parts or fabricated from plastic, brass or steel, as appropriate.

Interior

The interiors for the two passenger sections were also stereolithographed, complete with seats, toilets, sinks, baker heaters and partitions. Even the door handles on the interior doors were printed. The figures are from Preiser.

Finish

I airbrushed the body, roof and interior. Because the colour reference – a reporter’s impression of the colour – is so imprecise, I used Polly Scale Pacemaker Red straight from the bottle after comparing all my reds with my Turkish rouge sample.
The lettering was drawn in Corel Draw, working from some distant shots of CA coaches along with a crisper Grand Trunk car that seemed to exhibit similar lettering. From this lettering, I produced dry transfers, and installed on the letterboards and on the sides.
The lining was a nightmare. At first, I thought I would use a paint pen from Sharpie. While my initial tests on scraps went fine, I couldn’t manage a consistently fine line on the model. I wound up stripping the side and taking a mulligan. The successful technique employs coloured pencil, and is convincingly understated. When you see photos of CA cars in the nineties and early 20th Century, the lining is barely there.
The interior was brush-painted with acrylics, as were the passengers.
The underframe is mostly wood, stained before installation. To obtain clean colour separations, the brakes and other parts were painted before installation.
The underframe along with the roof received the bulk of the weathering. On the underframe, this consists largely of dry-brushing and an overspray with the airbrush. On the roof, I used chalks and weathering powders.