Quite long transition curves are used on the prototype to reduce the level of sudden sideways acceleration forces that railroad cars (and particularly passenger cars carrying loosely seated or standing passengers) would experience if the track went directly from straight to a curve radius and vice versa. Visually, such transitions make real prototype track, and of course then the trains themselves, appear to more naturally "flow" in and out of curves.

However, the purpose of transition curves on a model railroad is twofold, mainly because model railroads tend to have to use much sharper radius curves than the prototype.

First, as per the examples of the model ore train below, they can provide the similar to the prototype cosmetic realism appearance of "flowing" trains and long cars staying smoothly well aligned, even when moving on and off the relatively extremely sharp model railroad curves.

Second, model transition curves also importantly work functionally in that they keep working model couplers and coach diaphragms within a much safer working degree of sideways movement, and thus prevent forces which may otherwise cause derailments.

Here are examples from two different viewpoints of long 85 ft passenger cars running through the 24" long computer generated transition curves from a 48" radius end section of our new, under construction, test and demonstration layout.

Note that now even the long car ends line up to within 1/8" all the way throughout the transition from curve to straight, with just a very limited and smooth relative sideways motion and no wobble.

The scientific theory behind a transition curve is to make the curve radius change linearly with distance. That is to say that the radius increases or decreases at a constant steady rate between the straight and fixed radius portions.

While this sounds very technical, the shape of such a transition curve is just a chunk of the much more familiar "spiral" curve that you see on something like a watch or clock spring. Most of us get a similar effect while driving a car, when we slowly and steadily turn the steering wheel from straight to the position needed to gently round a bend. Sensibly doing that, instead of suddenly jerking the wheel all the way around to match the curve, and violently throwing our passengers (and their drinks) across the seats.

Since this part of the demonstration layout is supposed to represent a class 1 railroad's 4 track main line, I made use of a model railroad track planning computer program to draw out templates for the transition curves. Apart from saving me a lot of maths and drawing work, the program ensured that each transition curve was as accurately formed as possible, just as it would be if laid by a track gang on the prototype. Also the program automatically kept the 4 tracks aligned by making each transition curve slightly different so that each track stayed at proper 2" spacing larger than the one inside of it.

There is also a frequently mentioned "bent stick method" of using a piece of springy wood or "the naturally free curving shape" of flex track, to quickly create a workable compromise for a transition curve. Either method will give a moderately good "cosmetic" appearance curve but is really only a way to approximately "smooth out" the transition, rather than more precisely optimize the car end and coupler swing. If your own layout is close to the limit for coupler swing, or you have passenger trains, parallel tracks, or like me, just lots of different curves to plot, then I would personally recommend the extra moderate investment of track planning software. It's definitely one of those situations where technology actually helps the Model Railroading experience.

Just for the record, the track planning software I used was called 3rd Planit. I think it is still available, although my version is many years old. There is at least one other current product called CADRail, which I believe has similar features. For any such software, you don't have to calculate the easements. The software fits them in automatically as you draw your plan and plots the corrected track positions, including any sideways offsets of the straight track sections that are caused if you want to leave the fixed radii the same.

Finally, I should mention that my most desired requirement was to optimize (minimize) the relative end movement of long passenger cars going from straight to curved track. Since an 85 ft car is already about 12" long, it is fairly clear that a transition curve 12" long or less is not going to have much, if any, effect. So I chose what I think was the "long" transition option in the software, which turned out to be 24". The results, as you can see from the video clips, are quite presentable and realistic. Ideally, the the car ends will stay closer together if the transition curve is even longer, or as long as possible in the space available. But this was a first test, and I didn't want to have all transition and no fixed radius in my end curves. If I ever want to relay the track, I think next time, I'd try 36" long transitions, but like most of you, I won't be relaying this 4 scale miles of track anytime soon :).