November 8, 2009

Selective Compression of the Potomac Creek Bridge

I spent the weekend getting the dimensions of the Potomac Creek Bridge to my liking. A problem you sometimes encounter in building model railroads structures when you can't build a full sized replica is, how do you scale it down to make it fit and still look like a believable structure. Model railroaders call this process selective compression.

This drawing is from Haupt's book on military bridges. It is a good drawing of the arch-truss bridge similar to Potomac Creek, though there are some differences.

No scale is given for this bridge, but using the track spacing at 4' 8.5" yields a total span of about 160 feet. In O Scale this bridge should be about 40 inches long. This is actually longer than the arch spans at the bridge at Potomac Creek. But it does show the general construction details. It is also a through truss while the Potomac Creek Bridge is a deck truss (i.e. the tracks ride on top of the bridge instead of inside the truss "tube.") It's hard to believe that wood trusses can be so long and still carry the weight of a train.

So how big should I make both the overall bridge and the individual members? As I mentioned in an earlier post the direct scale down by 50 percent resulted in a bridge that was too spindly (though it would work for HO Scale). To resolve this dilemma I built two more versions of the truss.

Using the guidance from Haupt's book, I built a 48 inch long truss using chords and braces at 0.064 by 0.25 inches as this works out to 3 by 12 lumber in scale. That ended up looking like a grade school Popsicile bridge. This shot shows the laser-cut 0.25 inch pieces being assembled on the workbench. The pins help hold it place while the glue dries. I tried making pins from brass rod, homemade brass washers and solder, but that was a lot of extra work that is not needed for strength. I'll just use NBWs on the final model.

I rechecked Haupt's book and sure enough, he says the braces should be 2 by 10 and not 3 by 12 lumber. Thus the braces should be slightly more slender than the chords. I redrew the model bridge and decided that even that was a bit too wide in my reduced to fit span.

So the compromise I ended up with is selectively compressed in a couple different ways. The chords remain at 0.064 by 0.25 inches (3 by 12), but the braces are reduced in width to 0.186 inches which works out to about 3 by 9-by lumber. I also changed the number of stations in the bridge. Instead of 16 as in the prototype, I now have 10. In this case "stations" is used in the civil engineering sense where each vertical brace is a station. This shot shows all three trusses next to each other. The middle one is the one I am going with.

The bottom photo shows the final dimensions in the intended location, though it stills needs to be fitted to the scene.

One of the disconcerting things I noticed in reading the Googlebooks version of Haupt's Military Bridges book is that the plates cited in the text do not match the plates in the appendix. The scanned Googlebook also has some plates missing. For example, Plate 3 is in the paper copy of the book, but not in the Google scan.

Also, I noticed one other perplexing trait of this book. In the text, plates are cited with Roman Numerals, but in the appendix the plates are listed with Hindu-Arabic numbers. I noticed that Official Military Atlas to the Civil War does something similar but opposite. In that period book of maps, the plates are listed with Roman Numerals while the index is hindu-arabic. In any case, I find it weird at best and confusing at worst.


  1. Bernie,
    Are you using the laser to cut vertical and diagonal bracing from thin plywood as well as burn the hole for the pins?
    Is the span of the 12 vs 16 stations the distance between pier centers?
    The 'Haupt Truss' with pins seems like a arch reinforced 'lattice truss' rather than a Howe Truss. Your jig and fixtures look very helpful.

  2. Yes, the laser is cutting all the pieces to exact length and "drilling" all the holes precisely. Thus all the parts fit and are interchangeable, just as Haupt describes in his book. It took about 2.5 hours to glue up one side of the truss. I use the pins as a temporary holder. Once the glue sets, I pull the pin and reuse it to glue the next section.

    There will be 12 vs 16 stations between pier centers.

    You are correct, it is a lattice truss with an arch. Haupt says the arch supports the dead load, while the truss prevents the arch from "losing its figure under live load."

  3. One other point, the braces are 1/16 basswood, while the chords and arches will be laser cut 1/8 plywood.

  4. You do some inspiring work! I would like to do a similar project in O scale but I would build scenes of the Central Pacific constructing across California and Nevada. Only problem is I need an SMR trains CPRR Atlantic in 2 rail. Any of your buddies have one they would part with?


  5. Jason,
    Check with SMR Trains. He is preparing a second run of the Masons and could offer you an undecorated model. Then you could paint it to fit your RR. He is responsive to requests from his customers.

  6. Good job doing and describing the selective compression trade-offs...