Replicating a Distant Structure, the story of my attempt to create a model of the Deshler, Ohio, Baltimore & Ohio, passenger station.

It is funny how circumstances can conspire to generate modeling projects. This summer of ‘stay at home’ left me thousands of miles from my ongoing projects, my tools, and my modeling supplies. I didn’t want to try to duplicate everything but didn’t want to sit entirely idle. Searching for something to fill the time led me to Rail Cams. I’d seen the Trains Magazine Rochelle, Illinois camera before, but had never explored further. Boredom led me to YouTube and the cameras at Deshler, Ohio, Fort Madison, Iowa, and Santa Fe Junction in Kansas City, Missouri. While Fort Madison and Santa Fe Junction are busier, at Deshler, the buildings in the frame kept drawing me back.

The passenger station, and an old control tower now serving as storage for track crews, both loom prominently in the Deshler cam view. The tower is a fairly standard rectangle with some interesting details. Deshler’s old B&O station also appears pretty standard at first glance. It is also kind of beaten up, which makes it more intriguing. Pure idle curiosity led me to do a Google Earth street view to take a run around the nearest streets to look at the building from all sides. Given my usual luck, I wasn’t surprised that none of the street view images came anywhere near the station. I could only get distant views. That digital excursion further piqued my interest. I started looking for photos online.

Fairly quickly, I found Figure 2, showing that the two main lines intersecting in Deshler cross at 71 and 109 degrees. The photo also showed that the station was built with the walls facing the tracks, parallel to the tracks. That was enough to get me hooked. While I like constructing or detailing cars and locomotives, I love modeling oddball buildings. Now that I knew the station wasn’t square, I wanted it in three dimensions, in HO scale.

Now, I’ve never, to my knowledge, been within a thousand miles of Deshler, and this summer I wasn’t likely to make the journey to measure and photograph the station. That merely made figuring out how to generate a reasonable representation of the station more entertaining. Could I, using only material freely available online, come up with enough data to produce a decent model of that curious building?

In a way, I’d already started that process when I gathered photos and looked over the station on street view. Now, in order to try to build the structure, I needed to study those photos in detail. This is the most dangerous part of the project. It is way too easy to get caught in a state I call ‘research paralysis’. Sure, I had to spend some time gathering and reviewing photos. But I had to review them with the goal of finding some way to begin sizing the building.

The photo that gave me the angle of the tracks also gave me a known dimension-4’8”. Well, given the blur that occurred once I enlarged the photo to usable scale, not only did the omitted half inch go away, I called the tracks five feet and kept the slop in that decision in mind as I laid out the structure. I had a clear shot of the station roof, and using that five foot number, I determined that the roof was about 32′ across. Looking at the side views of the building I’d found, I decided that the roof overhang was somewhere in the area of four feet. If I went with four feet, deducting the overhangs would leave me a nice 24 foot wide building. Deciding the wings were 24 feet wide allowed me to estimate their length at 36 to 40 feet.

The next step was to try to confirm those numbers. I made my first stab at sizing the walls by trying to fit the windows and doors to them. There are some fairly standard sizes that appear again and again, 6’8” doors, for example, or windows that are 39” or 42” wide, and belt lines that are between three and four feet off the ground, or steps that are around eight inches in rise. Also often useful is the size of typical bricks, which with mortar, are about 3x4x8. Most of my photos of the station blurred when enlarged, making brick counting a problem. I could tell there was some fancy brick work in the building, but I couldn’t count rows and columns.

I found an image of the station showing people around when it was still in use. Based on apparent heights of pedestrians, I had a bunch of 9 over 9 windows, that appeared to be more than 6 feet tall. If the windows were that tall, they were also more than 3 feet wide. I also had a shot showing the original loading door, which I guessed would be 8 feet tall and wide, or maybe a little less.

Between the overhead shot using track dimensions, a little counting of fuzzy brick, and some estimates based on guesstimates of window sizes, I began drawing walls and floor plans. Was my first drawing going to be my last? Highly unlikely. However, beginning to draw up scale plans seemed the best way to discover what I’d guessed wrong, or otherwise overlooked.

After laying out estimated wall lengths and heights, I laid out estimated windows and doors. An eight foot wide (guesstimated) loading door allowed me to estimate window widths and spacing between the windows. Laying out the windows and doors with appropriate separations forced me to lengthen the main sides of the station.

The key wasn’t whether I had the exact right size for the windows, the key was having the windows in proportion, nor only for height to width but also with respect to positioning in the wall. In that first pass at drawing, the sizes for the windows provided me with a wall height.

Since the station wings appear to be symmetrical around a center line drawn through the middle of that big front window, I made certain my drawing was square by folding the drawing of the floor plan along the center line to confirm that the wings were exactly the same size and shape. Once I had confirmed that the drawing was square and equal, each change to the size of the structure could be made by adjusting the wings equally.

The foyer was a key design feature. When I began to really work out its dimensions, I was again and again forced to adjust various dimensions. Initially, I guessed that the three faces of the entry foyer were all the same size. The roof above them suggests so, as do the various window treatments. My first guess at the length of these walls was nine feet. That lasted until I tried to fit two foot wide columns on that wall and draw the big square window between them. The positioning of the windows and doors demanded resizing those walls to eleven feet wide. Which required a commensurate lengthening of the side walls to allow the windows in the side walls to be positioned a proper distance from the corners of the walls containing the main entry doors.

Deciding exactly where to place those three walls was a problem that would haunt me. It is evident that the walls containing the doors aren’t parallel to the main sides or the tracks. My first stab a placing those walls had window facing the diamond a foot outside the point where the extended walls of the wings would intersect. That position appeared to give me the shallow slope I could see in some of the photos of the station.

I was quite proud of catching this issue early. I was wrong to celebrate my initial wall placement. After two or three rounds of adjusting walls to make the windows look like they were properly spaced, I tried standing up the station. Photocopying my drawings, I cut a floor and a ceiling out of foam core. The walls I cut from poster paper, mostly to keep them thin. Then I taped the walls to the faces of the foam core floor and ceiling. Two passes of this process got me a set of walls that seemed to work. This brought me to the point of trying to fit the roof.

The first question to answer when preparing to roof the station was how much did the eaves extend beyond the face of the building walls? Was my initial four foot dimension correct? Based on the height of the space above the windows, I still thought the eaves had to be four feet. While I ultimately determined the four foot number to be about right, I discovered that I had adjusted for the width of the gutter system that encircles the entire roof.

I sized the walls of the foyer at eleven feet wide and thirteen and half feet tall on my second stab at sizing the building, and never changed that number. I was happy with those walls right up until I tried to build the octagonal roof over them and tie it to the roof over the wings. Then I learned that the foyer wall facing the diamond had to move back until it was a foot behind the projected intersection of the side walls rather than a foot ahead of it. I also had to shorten the difference between the short walls and the tall ones to about two feet total.

Interestingly, this two foot movement of the foyer wall didn’t directly translate to the side walls, Those were only going to have to get another foot longer to make room for the windows, bringing the main walls to about forty two feet long, which was as sensible as the twenty four foot width of the wings. That was a simple change. Yeah, right. Moving the walls means the floor and ceiling pieces have to get a foot longer. The roof sections over the wings have to get a foot longer… And…

All of which could have been caught by figuring out that the octagonal roof actually has exactly the same pitch as the roof on the wings. Once that becomes clear, it is possible to calculate the point where the eave has to be to place the point at the peak of the octagon in line with the ridge lines of the side roofs. The roof then will place the foyer walls, and all that long ago high school geometry might even help design that octagonal roof.

Of course, I made a huge mistake on my first run through. I looked at the short guy passing through the doorway in an old photo and convinced myself that the door to the station was more than 6’8” tall. I decided it was 7′ 6” went with that dimension all the way through constructing a full paper and cardstock model to prove the concept. The door assumption actually held good until I tried to convince myself that the windows were as wide and tall as I’d drawn them. Suddenly my side walls were looking S scale tall and somewhat over length for HO as well. Going over the photos of the station again brought me back to that door. Remembering that the typical size was an old standard, I made it 6’8”.

Once I fixed on the 6′ 8” door, major building dimensions began to make good sense. The wall height became about 11’8”, including the 6” to 8” of exposed concrete at the base of the brick, which also seemed to match the height of the step up to the doors. The shorter and narrower windows produced track sides and rear sides of the main wings that were the same length, which made perfect sense.

Another detail that finally became clear in the third attempt at drawing the station was the nature of the join between the section of the octagon that held the main doors and the window facing the diamond. The short stub walls connecting those three short walls to the wings weren’t perpendicular to the wing. Instead, they were short segments of additional walls of the octagon. This discovery further supported my late conclusion that the entire roof at the lobby area is merely a complete octagon poking up through and between the two pitched roofs on the wings of the station.

Figures 7 and 8 below clarify that relationship.

At last, I had a good working drawing. Now for the details. Like many brick buildings from the late 1800s and early 1900s, the Deshler Station had some fancy details. Some soldier courses, some raised brick rows and so on. Those would be difficult to get right, but not impossible, if I could find the right varieties of sheet brick. The problem was that the doors and windows on that foyer weren’t something I was going to be able to kit bash from anything on the market. Hand constructing them from stock styrene shapes was going to be a bear. I’d been through that aggravation when I built Bangor’s Union Station. Fortunately, or so I thought, the modeling world has evolved since I built Bangor’s station windows. Three dimensional printing was available and might be an answer. I’d bought and used some parts from a company called Shapeways that was set up to print items as drawn by their customers. Could I find a way to draw what I needed without purchasing an expensive CAD program?

My CAD search quickly led me to TinkerCAD. Free, browser based and seemingly not too difficult to use. Rather than drawing lines, the system was based on placing and stretching or warping shapes, and then adding or subtracting other shapes to devise the final desired item. A bunch of long narrow boxes would form window muntins. Windows then were just lots of boxes of different shapes all tied together. How tough could this be?

First of all, picking shapes and removing parts of them by turning other shapes into ‘holes’ was a different way of thinking. I was about halfway into drawing my first window, in HO scale (and the hard way), when I discovered that someone had produced and made available a section of brick wall. The bricks were deliberately irregular, and irregularly spaced, but all that was adjustable. Even the bricks themselves could be resized.

That was my ‘Hey, look, a squirrel’ moment.

Everyone who reads this is probably familiar with scope creep. Well, when I found that piece of brick wall, scope creep grabbed my project in a big way. Over the ensuing month, I spent most of every day engrossed in TinkerCAD. To begin with, I ran tutorials and quickly decided to toss away my partial window. There are dozens, if not hundreds of tutorials out there. The ones I found most straight forward and understandable were by a guy presenting himself as PromoAmbitions. I book marked several of his videos and returned to them multiple times to gain better understanding of some of the TinkerCAD features. Time well spent, for certain.

I learned a lot from the tutorials. There were a couple of aspects or tricks I discovered in TinkerCAD that aren’t covered in anything I saw. First, TinkerCAD gets slow and will bog down completely if you ask it to merge too many parts at once. In my first attempt to build my station walls, I made the little sample of brick into one big HO scale brick wall and inserted and soldier courses that I knew appeared everywhere. Then I just cut down, poked holes, and inserted more soldier courses and raised bricks to turn the big length of wall into the shorter walls I needed.

I shortly learned that TinkerCAD carries every single bit of the entire wall along the road to the smaller wall, and it sees every brick as an individual construct. By the time I finished building my walls, I couldn’t keep all of them in a single project, much less stand them up a look at the station. It was back to the drawing board to make short chunks of wall that had precisely matched brick sizes while not being much more than the size of the walls I needed. When making shorter walls, I learned it was important to calculate even numbers of courses and columns. You can only have full rows and courses, but can have any dimension for the overall wall width and height. Trying to produce half columns or part rows distorts the brick sizes so that the bricks don’t line up at the corners. Make the wall a row or column oversize and cut it down. Figure 4 shows an incompletely rendered project. Figure 5, when enlarged, shows that the bricks in walls of different heights still line up perfectly.

One of the options in the TinkerCAD menu is to zoom to the chosen element, whether a box, cone or whatever. Seemed pretty useful, but if you’re looking at an HO scale 8 foot door frame and want to see whether you’ve brought the corner together with the header, zooming in on the frame doesn’t get you close enough to guarantee a perfect joint. Further, you have limited ability to zoom further after focusing on the wall. Once I found that my connections weren’t as precise as I wanted or needed, I also discovered a way to get the tight zoom required to neaten up the details: Whenever I wanted a close look at a corner joint, a window frame, or whatever, I dropped a box beside the point of interest and made the box’s largest dimension no more than 2mm. Zooming in on the box achieved a much tighter zoom than any other method and made that 2mm feature fill the screen, Once I was zoomed in tight, I could scroll across the drawing to the point I wanted to review. Whenever I waited and zoomed like this, I always found that pieces of the project weren’t tightly aligned. Eventually I learned to just keep a little cone handy and drag it close to what I wanted to see.

A third tip involves the TinkerCAD snap feature that controls movement when you drag objects together. 1.0mm is the standard snap distance in TinkerCAD. Every time you load the program, it defaults to this. Since an inch is less than 0.30mm, many of your dimensions will end up less than 1.0mm, making a 1.0mm snap an aggravation. If you’re working in HO you’ll almost certainly want to set the ‘snap to’ dimension to 0.1mm, and at times you’ll want to turn it off entirely.

Figure 5. The top of the front window and the tops of the windows above the doors are at the same height. That feature ultimately defined the wall widths and the additional height of the front walls versus the walls of the remainder of the station. The reddish area to the left is something that didn’t fully load, and the open space behind the doorway is showing a spot where a wall didn’t load at all.

Figure 7. The octagonal roof in proper position relative to the roofs of the wings.

Would I have gone down this particular rabbit hole if I’d been home this summer? Probably not. I’d have continued to struggle with traditional means of constructing the projects I had in progress last fall. Do I regret diving into this? Not at all. The possibilities are almost endless. In between working on the station and another building, I’ve made a harp switch stand, the interior details (and the side juts) for one of the BAR troop sleeper cabooses, and a number of detail pieces for my Monson slate quarry portion of my Monson RR. Each and every one of those items is something I had become stuck on.

What unique item is lurking in the back of your mental ‘I wish I could figure out how to make this’ list? TinkerCAD and 3D printing may be your solution. Oh, and as proof that TinkerCAD will handle some brick buildings, below are a couple of shots of my fully drawn and standing Winterport water works building. It still needs a chimney and probably a vent stack, but once I draw those, I’ll lay it down flat and sprue it together and ship it off for printing. I had taken pictures of the structure that included a scale so there should have been none of the guess work involved in the Deshler Station. In the end though, I scaled and laid out this structure by counting bricks.