Working on the smaller details of the box car gives me a taste of what other crafts must be like. I feel like a jewelry maker because the work is delicate and fiddly but only in the sense that I am not used to it. These details require as much, if not more time, than the major construction.

Freight Car Jewelry
The door tracks were an exercise in heading scratching as I needed multiple copies of the brackets, all to the same shape and dimension. I have written about this before so will move on.


Details like the end ladders were straightforward thanks to prior experience and a solid jig that holds everything together while I solder the rungs to the stiles. Once soldered, they are surprisingly robust and clean up nicely.


The most delicate detail is this assembly on the bottom of the draft sill extension. They are part of the cushioned draft gear system and I had to stare at my reference photos for a long time before figuring them out.

Cushion bracket


Cushion bracket_00

My objective was to eliminate the clunky coupler box and mounting screw we typically use. The screw to remove the couplers is out of sight in this view but accessible. This assembly is easily seen on the full-size car and I will go into greater depth about its construction in TMC, Vol 12 but, for now, I want to address a concern many will voice about how a scale underframe and coupler mounting will never work for model railroad curves.

I was concerned about whether the car, as built, would negotiate my super tight 48 inch radius curve. Remember that I work in quarter-inch scale and 48 inches is the rough equivalent of a 24-26 inch radius in HO. My fears were laid to rest the first time I mounted the trucks and gave the car a test drive on the layout. I discovered there was nothing to worry about, at least in this case. As you can see in this photo below, taken at the end of the layout where the staging cassette normally goes, the inside faces of the wheels do not come close to rubbing against the center sill or the sides of the bracket, nor do the wheels come close to rubbing against the side sills, even though the car is on the sharpest kink in the curve.

Close but no cigar!

Close but no cigar!

Notice too that there is plenty of clearance on the top of the flanges for the I beam floor stringers. I do need to add a thin spacer between the truck and the body bolster because the car sits too low by just a tiny amount. This will help the top of the truck to clear the bolster as there is some minor contact especially on the banked curve where the car wants to lean to the center of the curve.

Coupler Swing
Another convention we deal with is the idea that you need a huge amount of coupler side-to-side swing. On full-size cars the coupler swing arc is only a few inches from side to side. Long cars like 86-foot HyCubes or TOFC flats may have a greater arc due to their length. In both cases however, the prototype has nothing like the gross side-to-side swing arc of the typical model. I am thinking of standard common-carrier equipment with that statement. I do understand that interurban rolling stock had specialized coupler mounts with greater swing arcs.


In the photos, you can see the narrow width of the box car center sill. It is a scale fifteen inches wide between the inside faces. In this photo You can see how sharp my curve is by the severe offset that the cars have to each other. Again, 48 inches is stupid tight for quarter-inch scale, being more suitable for industrial trackage, narrow gauge or traction. I could not get the camera positioned exactly as I wanted but notice in the photo how the shanks of the couplers are not that far off center, with the bulk of the offset occurring at the knuckles. These are both fifty-foot cars, so the offset is balanced between them. As we all know the problems come with cars of unequal length, and that is my experience too on this curve.

(As a side note, you can see that Prototcraft couplers will interlock with Kadee and even Atlas as shown here. They won’t couple hands free but they do hold when coupled artificially. I am not advocating a mix and match approach to coupler choices, just pointing out an interesting fact.)

On a full-size curve this tight, the crew would take it at walking speeds or less. If the locomotive could not handle the radius, they would use others cars as a handle, a trackmobile or other wheeled vehicle, maybe even winch them in and out with a car puller cable system if there was one. In essence, whatever is required to get the car where it had to be.

Myths have power
Many folks, myself among them, still carry a lot of baggage from the toy train era. The myth that you need a massive degree of coupler swing and truck swivel motion for satisfying operation still exerts its influence to a surprising degree. It is especially bad in quarter-inch scale as the three-rail mind set drives the manufacturing of most commercial products. Whether they are billed as three-rail or two, the body bolster design is exactly the same.

The thought I want to leave you with is to experiment for yourself. As I gain more experience with rolling stock, I agree more and more with the late Bill Clouser, that when you build things like the prototype did, they will operate like the prototype did. Is that a universal truism? Who really knows? I do know that it is working on my layout and that your situation will be different. Proceed accordingly.



  1. Simon

    Lovely, interesting, post, Mike.

    One thing I have learned is that engineering tolerances need to be equal, for finescale that means equally tight, for standards aimed at coping with a train set on the carpet, that means equally loose. If there is a lot of sideplay in the wheelsets, then there needs to be a lot of sideplay in the coupler boxes.

    Closing up these tolerances needs to be systematic and matched across the board. In your case, you may well obtain better running round your 48 inch radius curve than you would round the same radius using 3-rail track standards.

    Bill Clouser was almost spot on, but there are some things which cannot be directly scaled (at least not practicably) and scaling directly from the prototype can lead to problems, as some Scale7 modellers have discovered over here.


  2. mike

    I agree there are practical limitations with models. I do contend that we can get far closer to the prototype than the mass market mentality would have us believe (hence the finescale movement). It really depends on what one wants to accomplish and the degree of slop you are willing to live with.


  3. Simon

    I agree: less slop = better running.