Twist and Cant - How much is acceptable?

[Will L]

While I was doing a session on CSBs at a recent meeting of the Crewe area group, the question of cant came up. While I know that there can be an element of “empty, uncritical thought or talk” (thank you Wikipedia) which can work it’s way into the questions about the right and wrong way to suspend your locos, the cant in question actually related to the difference in height between parallel rail heads. This occurs intentionally, when the track is super elevated round a curve, or unintentionally, when we fail to get the track as flat as we intended. As we build most of our track to be as flat as possible, for cant to occur there must be a section of track that is twisted so the difference between the railhead heights is changing over a distance. It is this twist that is a distinct challenge for our model suspension, and an even greater challenge for stock that is totally rigid. It is the key reason why the first thing that tends to happen when you lay a super elevated curve is that some of your stock comes off trying to go round the corner, usually the stock with the longest wheelbase.

At out meeting it was suggested, by the track gaffer on Knutsford East, that in the same way as we may specify a minimum radius, there should also be a figure for the maximum amount of twist our locos would tolerate. Note that it is not the absolute amount of cant that we are worrying about here, as any constant value which is visually acceptable should cause us little difficulty. It is the rate of change that is important, and it is measured in turns of a gradient between the two parallel railheads.

I must admit I hadn’t given it much constructive thought until that moment. I do know that I’m constructing loco’s which by designed should tolerate no more than 1mm deflection between the lowest and highest wheels. So while I would expect my locos to deal comfortably with some twist, I haven’t any real idea how much, beyond suspecting that problems would emerge well before you get to the design limit of a full 1mm over the coupled wheelbase (say 1 in 60).

Having accepted the concept that one should specify a level of twist stock should be able to tolerate, it seems obvious that the next step would be to build a bit of test track that contains a section that twists at the maximum design level. However, before I do that, I need to decide what the maximum twist or cant gradient we should expect our locos and stock to navigate is.

To this end I (re) read digest 21.1 Designing the Whole System - The measurement and setting-out of curves particularly section 10 onwards. This gives you a timely reminder that transition curves exist primarily so that cant can be introduced progressively as the radius of curvature becomes shaper, and that on the prototype the maximum permitted cant gradient is 1 in 400. But then it also points out that because the physics doesn’t scale, cant deficiency is all but unknown in 4mm scale, so we only super elevate our track to emulate the original, probably to the detriment of running. Not only that but given our foreshortened distances there is likely to be insufficient room to include a long enough transition curve to permit prototypical amounts of cant combined with prototypical levels of cant gradient.

I can’t say that helped much, so perhaps I will have to set up track with a variety of cant gradients and see what happens. I could also go back the Knutsford Track Gaffer and ask him what cant gradient he designed in, that should amuse him.

What do you lot think?

[andrew jukes]

I suspect twist as a result of cant change in a transition is never going to be as big a worry as, for example, a dipped rail joint in one rail only.

Typical parameters I'm using are to cant curves by from 1 in 100 to 1 in 40 (depending on likely speed and nature of traffic) and to have transitions of typically 600mm-800mm in length. So the worst case is likely to involve 0.5mm of twist spread over 600mm, or 1 in 1200. No problem, compared with, say, a 0.2mm dipped joint for a 50mm wheelbase vehicle. I've lots of track laid but no serious running yet, so the theory is still to be properly tested.

The chosen cant is enough to make a big difference to how a train looks in a curve without (hopefully) leading to the problems one would expect with the wheelsets bearing too heavily on the inner rail. No way would I ever use anything like the prototype's max cant of 1 in 10 on a 4mm scale model.

Andrew Jukes