... 1:10 crossing, the knuckle gap is 2.5/32" (i.e. the knuckle gap is wider than the normal flangeway gap of 1.3/4 "). Here's a pic showing an almost as-new bullhead crossing. You can see from the wear pattern on the knuckle radius how the twist in the rail has changed the angle of the rail top. Notice also the heavy bolted slab& bracket "A" chair construction, holding the wing rail down level. Thanks to Mick Nicholson for the pic: 2_061131_520000000.jpg© Mick Nicholson For the vee rails, the rails are vertical where they are spliced together at the nose. A twist is made in the rails to return them to the usual 1:20 cant -- the twist is usually between the B and C chairs in the first diagram above. It's not an easy twist to replicate, because the gauge face of the rail must stay in a straight line as the rail foot is bent outwards. From about 1970 to 2000, flat-bottom turnouts in the UK were ...

... of the chairs. Obviously on the prototype the rail can't drop below the other rails if it is supported in a cast-iron chair which has the same base thickness. The result without the correct twist is that the far end of the wing rail tends to kick up, and must be held down in the chairs by means of the keys, spacer blocks and bolts. If you look at the wear pattern on this knuckle bend, you can see how it changes as the rail table is twisted through the bend: 2_061131_520000000.jpg photo thanks to Mick Nicholson On the prototype the rail twisted on a hefty machine in the crossing shop. Making a similar combined bend/twist in inclined model rail is very tricky to get right, which is why for traditional riveted and copper-clad construction many modellers prefer to keep the rail vertical. To allow for the wheel coning, the point rails should be reduced in height a fraction at the tip, so that the wheel can roll smoothly off the point rail onto the wing rail. And if I can ...