How do trains stay on track? (2009) [pdf] | Hacker News Reader

How do trains stay on track? (2009) [pdf]

materials.mcmaster.ca

139 pointsscrrr29 comments

It's similar to the principle that keeps flat belts from slipping off their pulleys...

http://woodgears.ca/bandsaw/crowned_pulleys.html

...even in extreme situations such as this:

http://en.wikipedia.org/wiki/File:Transmissionsriemen.jpg

I wonder at what point you start getting a convex pulley having a (local) optimum in terms of how likely the belt is to stay on the pulley?

For low speed / low friction between belt and pulley, you'll just pull the belt off of the pulley if the pulley is convex.

And for high speed / high friction you'll end up with the above effect.

So I wonder: where does it transition?

If the belt is rubber, the friction is high enough, even when the system is starting from zero. (Assuming pulleys shaped like the ones linked above, of course.)

Is that the case even for a (mythical and completely useless) frictionless pulley?

Whether a frictionless pulley is useless or not depends largely upon whether you are using it to change the direction of the line, adding tension to it, or using it to perform work. Such a pulley is only useless for the latter, such as a belt turning a drum, but a frictionless block and tackle or tensioning pulley would be incredibly useful.

Good point.

So my question isn't as odd as it might first appear then.

additionally, here is Richard Feynman explaining it: https://www.youtube.com/watch?v=y7h4OtFDnYE&index=7&list=PL0...

Love to hear Richard Feynman explain anything.

Right? I wish he was alive today so he could do a ton of YouTube videos explaining simple counterintuitive concepts like this.

Here's him explaining how computers work (first 35 minutes): https://www.youtube.com/watch?v=EKWGGDXe5MA

That whole interview is so great.

A great video, and I noticed the actual linked article here suspiciously follows the exact same structure and examples that Feynman does in his explanation. It would have been nice to acknowledge the fact that the author was probably watching Feynman explain it while writing. But it seems to be a student fun journal, and has made a few hundred people think, so why get upset?

Maybe both Feynman and the author copied the explanation from a third source.

I wondered, how would this work with an axle-less low-floor bogie? The answer I found here[1] is — it doesn't. “… there is only a single series of trams that have these bogies, later versions returned to the traditional type with axles. … The wear problem is because those bogies don't seem to track as well, they go from left to right in the track”. The implication is that there were people designing rail vehicles who didn't know how they stay on the track.

[1] http://modeleng.proboards.com/thread/4521/chilled-iron-tram-...

There are more trams with axle-less bogies now.[1] The goal is to lower the floor height, eliminate steps, and improve wheelchair accessibility. It gets complicated. Some systems use big driven wheels and small trailing wheels. Some have a geared axle between the wheels. Siemens has a system with motors on each wheel, with the wheels locked together electronically. AEG tried something like that, with less success. Speeds of the more elaborate systems are rather low, below 70km/h (44mph). That appears to be a limit of the axle-less design.

[1] http://onlinepubs.trb.org/onlinepubs/tcrp/tcrp_rpt_02.pdf

Here's a video demonstration by Richard Hammond of Top Gear.

https://www.youtube.com/watch?v=OEo-hQbyy30

That is really good, and also a massive reminder of how meaningless and inconsequential my work is, compared to people doing serious electrical and mechanical engineering like that.

Also, a great video explaining the working of a differential:

https://youtu.be/K4JhruinbWc?t=110

anyone know if this was an intentional part of wheel/rail design from the start... or they originally intended it to work how we naïvely imagine and by accident it worked this way and was subsequently refined ?

I don't know much about it, but in looking at some of the first locomotives I saw this: https://upload.wikimedia.org/wikipedia/commons/e/e2/Richard_...

and this: https://en.wikipedia.org/wiki/Richard_Trevithick#/media/File...

That was probably a terrible screeching thing to witness going around bends, just being held in line by the inner lip of the rail.

In a similar note I wondered what the patent on the railroad ties on the Caltrain bed were (#5104039) and found this: https://www.google.com/patents/US5104039?dq=5104039&hl=en&sa... which has an interesting discussion about how the tracks manage to stay underneath the trains.

Does the rail have some flexibility to twist and meet the wheel cone at the right angle, and/or is it pitched to match the angle of the wheel surface? When I look at the tracks and this wheel shape, I see the wheel riding on a corner of the rail, which is obviously not the case.

The portion on the structural properties of wood a bit further down the PDF are just as interesting.

Yes and I really appreciated the reference to the USFS's Wood Handbook: Wood as an Engineering Material.

Available online:

http://www.woodweb.com/Resources/wood_eng_handbook/wood_hand...

Thanks; I hadn't heard of it. For what it's worth I looked up the source:

http://www.fpl.fs.fed.us/products/publications/specific_pub....

http://en.wikipedia.org/wiki/Hunting_oscillation

Edit https://cmnewsletter.wordpress.com/ typical, find the answer just after asking the question below.

This was brilliant but I'm really struggling to find an index page with the rest of them, I can find some of them via google directly.

Anyone figure it out?

I've been on and around many trains and yet I never quite understood the concept of the wheel flange and how track junctions worked until very recently, thanks to a kids show: chuggington.

I would love to try to simulate this with a computer program some day.

I really enjoy playing train simulator games, but as far as I understand, none of them actually simulate the physics of cone wheels and flanges.

Reply on news.ycombinator.com