Neil's News

Wheel Shapes

29 May 2003

Last year I saw an interesting activity at the National Railway Museum in York. There was a curved track and four sets of wheels. Each set of wheels had a different cross-section (as shown left). To no great surprise, when one rolled the flat wheels (A) down the track, they rolled perfectly straight and fell off at the first curve. The double-flanged wheels (B) also ran perfectly straight (which is surprising until one thinks about it). The outside-flanged wheels (C) actively veered off to one side or the other. The inside-flanged wheels (D) were self-centering and wobbled their way to the end of the track without derailing. About a month later I spotted the same activity in a different museum; it must be a stock item.

The beef I have with the exhibit is that it is being used to prove the wrong point. It is a fantastic demonstration of mechanical self-centering (as wheel D drifts left, the diameter of the left wheel increases which makes it travel further, forcing the axle to swerve right). This principle is used to keep belts on pulleys and keep aeroplanes from rolling. But the activity was claiming to demonstrate why railroad wheels are the shape they are. Anyone who has experienced the ear-splitting TTC ride between St. Andrew and King can attest to the fact that self-centering geometry is NOT what keeps the train on the tracks. Trains don't have conical wheels, they have flat wheels (which go straight) with sharp flanges (to keep them on track). You wouldn't want conical wheels wiggling their way down the track at 200kph.

This museum activity has been bugging me since it appears to be teaching a falsehood.

Update from NRM:

Thanks for your email and for your article about the NRM. We have had a look through you article which is good but there are a couple of factual errors. Firstly the track is no stright it is at an angle to match the coned shape of the wheel. And secondly the noise you can hear when trains are going around corners are caused either by the train going too fast into the corner and the flange is pressed against the track or the corner being very tight.

< Previous | Next >