superelevation, speed limits and pointwork

[Alister_G]

Hi all,

 

I'm still in the process of researching for my proposed OO model of Bakewell station and goods yard on the Peak District Midland line.

 

From photos I've seen, it looks like the trackwork through the station (which is built on a curve) is canted, but there is a crossover between the up and down line, and a turnout to the goods siding off the "high" side of the down line.

 

Reading this thread: http://www.rmweb.co....elevated-track/ it seems unlikely that the track would have been built like this, so am I wrong?

 

If it is canted, would they have reduced the elevation for the bit where the pointwork is, or would that effectively drop the speed limit for the whole section so be pointless (excuse the pun).

 

The photos I've been looking at are here:

 

Up Line:

 

http://www.disused-s...ll/index6.shtml

 

http://www.disused-s...l/index79.shtml

 

Down Line:

 

http://www.disused-s...l/index77.shtml

 

http://www.disused-s...l/index60.shtml

 

So, firstly, would you agree that the track appears to be canted, or am I being confused by the camera angles?

 

Secondly, does anyone know what the speed limit was through Bakewell, or where I might find that information, as that might give a clue as to whether it was canted or not.

 

Thirdly, if it is canted, how would I go about modelling the crossover and goods yard pointwork, or would you in fact bother? (I'm planning on using Peco code 100)

 

Would really appreciate some guidance on this,

 

Thanks.

[Poor Old Bruce]

1) The track would be canted.

 

2) I have had a quick look in the 1937 Sectional Appendix but it isn't very helpful as regards line speed limits - I think you needed to go on a course in those days to understand it. The next SA I have is 1969 after the line had closed. The information shoud be in that sort of publication if you can find one from the 1950s or 60s. You may have to go to Kew unless someone comes up with the goods on here.

 

3) The crossover would need to be reasonably flat, meaning that, if canted, the down line would have to be higher than the up. Personally I wouldn't bother with canted track, I think it is more trouble than it's worth.

[The Stationmaster]

The permissible speed depends on the relationship between the radius of the curve and the amount of lift in the cant - and nowadays it is very different frm the past for a variety of reasons.

 

On the GWR tables for cant a 40 chain radius curve needed 1 1/8th of an inch of cant to permit 30 mph and 4 1/4 inches of cant to permit 60 mph. The minimum radius on which cant was used was 5 chains and at that radius 4 1/4 inches of cant allowed amaximum speed of 20 mph. The highest amount of cant in the GWR 1936 table is 5 3/4 inches - which allowed 70mph on a 45 chain radius curve. The table goes out to curves of 200 chains radius.

(for younger readers there are 80 chains in a mile - so, for example, a 40 chain radius is a half mile = 880 yard radius or about 35feet radius in 4mm scale)

[Gruffalo]

Since your largest radius through the station is around 76" as The Stationmaster's data suggests you can ignore it. You could of course angulate your track bed so one platform is lower than the other and try twisting the trackbed a little but I doubt it would stay in the desired position without a lot of bracing and for what you would gain visually, I believe the lily doesn't need so much gilding and reliable running would be a higher priority in my mind!

[RBE]

I disagree. In model form curves are virtually always tighter than the real thing and even a bit of cant makes a very nice effect and well worth the effort.

 

Cav

[coachmann]

I introduced super elevation on my Greenfield layout in 2008 and did it again this year on the new Greenfield. A study of track photographs in steam days reveals that there was a lot going on in those days. A crossover between curved Up and down tracks is easy if you treat them both as lying on the same plane. This of course means the out track is much higher than the inner and calls for double the amount of packing on the track leading off the crossover.

 

A set of points on the ouside of a canted track leading to a siding would be canted too. This leaves the track leading to the siding pointing upwards of course, and so the track is then gently bent downwards again as it leads into the siding or yard.

 

At the real Greenfield, which I am modelling, there was a double junction on a canted section of curved mainline. As if this wasn't enough, the mainline was also on an incline. Simple arithmetic shows that the diamond left the mainline at a slightly lower point along the incline than the adjacent set of points. Smooth negotiation of the double junction was prioritised for the mainline and all the compromises were built into the branch leading off the junction. Trains negotiating similar junctions do not run as smoothly as one might think and it is interesting to watch the movement and swaying of coaches on old railway videos.

 

In model form, the slightest hint of cant is sufficient to capture the prototype running of trains on curves. In my opnion it is as important as anything else on the railway. Lay model track perfectly flat and it is not much of a step up from laying a kiddies roundy on the table.

[The Stationmaster]

I tend to agree with Coachmann on this. What modellers are trying to do is capture the 'impression' of the railway and thus often means resorting to all sorts of visual tricks ranging from selective compression of all sorts of things to 'selective exaggeration' of others. For instance in 4mm scale a 3inch cant lift is only 1mm while a lift of, say, an inch is hardly noticeable - so visual trickery has to come into play while almost none of us can hope to even halfway accurately portray prototype running line curvature.

 

Edit typo

[Armchair Modeller]

Occasionally, track laid flat will look as if it has a negative cant. This happened to me when I laid some main line double track on a tight left-hand curve, climbing up a gradient, with some level sidings on the outside of the curve. The optical illusion had me fooled for a while before I worked out what what the cause was. I relaid the main line with some super-elevation and everything looked fine. Since then, I have always used super-elevation where practicable.

[martin_wynne]

There is a difference between a running line junction, where both roads are running lines, and low speed crossover connections into a goods loop or yard. Regardless of any superelevation, such loops or yards are often at a lower level than the running lines. This uses gravity as a protection against vehicles running away, in addition to any trap/catch points.

 

In open country the running lines are usually individually canted, especially where tracks pass through bridge arches or other limited clearances. Unless they are widely separated this obviously makes it difficult to have a normal crossover between them. For areas of such connections, the cant will either be reduced to flat (with a corresponding speed limit) -- this is the most likely arrangement through stations where most trains will be stopping. Or gradients will be introduced to raise the outer (or lower the inner) track, so that the superelevation is in a single plane across all tracks. This means there is likely to be an adverse cant when negotiating crossover roads from one track to the other, and such movements will have a severe speed restriction. There will then be a gradient down into the goods loop, which should properly be spaced at least 10ft way from the adjacent running line (15ft-2in centres minimum).

 

In the case of running line junctions, special two-level chairs are combined with raked timbers to minimize the effects of adverse cant and reduce the need for speed restrictions.

 

Here is an old post of mine covering this in more detail:

 

How would the considerable drop to the yard be accommodated.....with a vertical transition curve?

 

If the tracks are at 20ft centres, and you have say a 1:8

crossover between them, the clear distance along the

track between the V-crossings will be about 60 to 70 feet.

So a drop of say 9" will require a gradient of only about

80:1, with some room for a short vertical curve at each

end. Such vertical curve will be put in by eye by the gang

when packing the tracks, and it definitely won't be a

mathematical transition if it leads only to a yard.

 

I came across a note in which I have written that superelevation is not applied through crossovers turnouts/etc. Am I correct in this statement?

 

For traditional track, no. If a curved track requires

superelevation (cant) in order to be comfortably

traversed at the line speed, and it includes a curved

turnout, then clearly the turnout must be superelevated

too. It's obviously not possible for the rails to suddenly

change from being superelevated, to being level for the

turnout, and then back to superelevated.

 

Nowadays you are generally correct, since track layouts

have been remodelled to simplify connections and move

them away from sharp superelevated curves.

 

In applying superelevation to double-track, you have two

choices. a) tracks independently banked on a level trackbed,

so that both low rails are at the same level, b ) a bank

across the entire trackbed so that all 4 rails are in the same

plane.

 

a) is the more usual arrangement, especially in open

country or where the cant is significant. It obviously

complicates the issue if you need a crossover.

 

b ) is used for more complex areas of pointwork where

they occur on a superelevated curve, since it simplifies

the connections considerably to have all the rails in one

plane. Especially where there are long timbers across

several tracks.

 

In either case, the connecting tracks crossing from one

line to another will be likely to have an adverse cant and

be subject to a low speed limit.

 

Here below is the text from the bible (in this case BRT3)

 

------------------------------

 

TWO-LEVEL JUNCTIONS -- BULLHEAD TRACK:

 

In an ordinary turnout from a Main Line, the turnout road is

almost invariably curved. (It is occasionally straight when it is

on the outside of a curved main). The turnout curve should,

therefore, have superelevation, but owing to the long timbers

which are necessary to pass under both roads, it is not possible

to provide this, and for this reason, among others, it is necessary

to impose a speed restriction on traffic passing over a turnout.

 

Where the main is curved and superelevated and the turnout is

on the outside of the main, and curved in the opposite direction,

a worse condition exists as the turnout will have superelevation

on the wrong rail, i.e. the inside rail will be higher than the outside.

 

This lack of superelevation, or superelevation on the wrong rail

of a turnout, besides causing discomfort to the passengers

travelling over it greatly increases the strain on the fastenings

of the track.

 

In the past efforts have been made to improve these con­ditions

by such means as adzing the chair seatings on the inside rail of

the turnout, and/or placing chair packings under the other chairs.

 

Neither of these methods is entirely satisfactory as adzing of chair

seatings cuts away the part of the timber that has been creosoted,

makes a pocket for water, and weakens the timber, and chair

packings tend to split, and they both decrease the length of chair

fastenings in the timber.

 

It is not possible to get over the difficulty completely by these

methods, and they are not easily applicable to a double-line

junction, where the outside road of the branch lines crosses both

of the main roads, and where normally long timbers supporting

all four roads are used for part of the junction.

 

A method of improving the superelevation of a junction is by

the use of what are known as "two-level chairs", and they are

especially useful in turnouts and crossover roads.

 

In an ordinary chair for switch and crossing work that supports

two running rails, the two rails are at the same level. In a two-

level chair, however, the depth of metal between one of the

rail seatings and the underside of the chair is increased so that

one rail is higher than the other.

 

Two-level chairs are made with the chair seatings increasing

in depth by increments of a fraction of an inch, generally 1/12th

of an inch or 1/16th of an inch. Taking the average chair spacing

in point and crossing work to be 30", it will be seen that by using

a succession of two-level chairs increasing in depth by 1/12" it

is possible to obtain a difference of level of 1" between two rails

in a distance of 12 x 30" = 30ft. If the chair seating depth increases

by 1/16" the distance necessary to obtain a difference in level of

1" is 16 x 30" = 40ft. (Usually termed a superelevation gradient

of 1" in 40ft, or 1 in 480.)

 

Thus with two-level chairs on the outside rail of a turnout, and

normal chairs on the inside it is possible to obtain a superelevation

for the turnout road of 1" in a distance of about 30ft or 40ft.

 

Two-level chairs in the 1/12" increment range are marked with

the letter T and a number indicating the number of twelfths of an

inch by which the base is thicker than normal. Those in the 1/16"

increment range are marked with the letter S and a number

indicating the number of sixteenths by which the base is thicker

than normal.

 

In a switch the switch rail and stock rail have to be at the same

level from the tip to the end of the straight planing. From this point

onwards two-level chairs can be used, and the level of the outside

switch rail of the turnout road raised above the adjacent stock rail.

 

In order to make the switch rail seat properly on the higher level

slide chairs, it is necessary to give a vertical set in the switch rail

equal to the gain in height at its heel divided by the length from

the end of the straight planing to the heel.

 

By giving a rake to the long timbers it is possible to use lower

two-level chairs than would otherwise be necessary. The timbers

are raked by an amount equal to half the superelevation required

for the turnout in the gauge, as explained below.

 

Assume that at the crossing of a turnout in a straight main road

2" superelevation is required on the turnout road. The long timbers

are then raked to give a 1" fall across the gauge. (This means

that, if normal chairs were used, there would be a supereleva­tion

of 1" on the main road and on the turnout road). Standard chairs

are then used for the main line rail opposite to the crossing.

 

As the two main line rails are to be level, and also the point

and splice rails have to be level at the crossing, chairs having

a base for both rails thicker than standard by an amount equal

to the rake, in this case 1", are used here.

 

Standard chairs are then used on the turnout rail opposite the

crossing. The superelevation on the turnout road is then equal

to the sum of the increased thickness of the base of the crossing

chairs, i.e. 1" in this case, and the rake of the timbers again 1",

and thus the required superelevation of 2" is obtained for the

turnout road.

 

It is not possible to use two-level chairs in a diamond-crossing,

as the two main line rails and the two branch line rails at a set

of obtuse crossings obviously have to be the same level for a

straight main line, or have the same superelevation as the main

in a curved main line.

 

At a junction in a straight main, where it is desired to superelevate

the branch line, it is therefore necessary to redesign the junction

to give as uniform a side thrust as possible. To this end the curve

through the obtuse crossings should be such as to give a cant-

deficiency as nearly as possible equal to that obtaining in the

turnout at the appropriate speed, subject always, in the case of

fixed obtuse crossings, to the following limitations of radius :-

 

Obtuse Crossing: Desirable Radius: Minimum Radius:

 

1 in 8 30 chains 25 chains

1 in 7.1/2 25 chains 20 chains

1 in 7 20 chains 15 chains

 

It is necessary for high speed junctions and often desirable

for other junctions to introduce transition curves within them.

This applies to an even greater extent to crossovers and

connections to and from loop lines since the change in direction

of thrust is therefore greatly eased. The fullest advantage of

two-level (thick based) chairs may be taken when they are

used with transition curves to ensure uniformity or minimum

rate of change of thrust. Provided that sufficient setting-out

information is given, there is no difficulty in installing transitional

junction work.

 

It is essential that point and crossing work with two-level

chairs should receive very careful maintenance. For instance

a joint which is 1/4" low in a standard turnout may not be very

serious, though of course it should have attention, but in

two-level point and crossing work, with a superelevation

gradient of between 1" in 30ft and 1" in 40ft, a 1/4" low

joint will con­siderably affect the running of traffic over it,

and may increase the superelevation gradient to as much

as 1" in 20ft. This applies especially where long timbers are

raked, and these must be firmly packed under each rail,

and the correct rake maintained.

 

Fig. 60 shews a thick-base chair compared with a standard

chair. The letter T and figure 12 on the thick-base chair

indicate that the base is 12/12th of an inch, i.e. 1" thicker

than the standard chair. Fig. 61, of two of the heel chairs

of D-type switches, shews how the switch rail is raised above

the stock rail.

 

TWO-LEVEL JUNCTIONS -- FLAT-BOTTOM TRACK:

 

The use of two-level switches and of baseplates with

thickened rail seatings has been continued in the case

of FB material. While in the main following BH practice,

one important change has been made for the FB designs.

 

A decimal system is used for the increases in rail seating

thickness in place of the fractional system of increments.

The minimum increase in thickness between consecutive

baseplates is 0.02" (20 thou). This decimal system enables

a much wider range of cant gradients to be obtained.

Additional identification markings are of course required

on baseplates which are thicker than standard.

 

The marking 2/L is used to indicate the association with

two-level switches, followed by a number which gives the

additional thickness of rail seating in hundredths of an inch,

e.g. ST2/L 50 would denote an ST baseplate with seating

thickened by 1/2".

-------------------------

 

Bear in mind that there are no hard and fast rules -- different companies did things differently, no two sites are identical, and whatever you decide, someone can be relied on to produce a photograph showing the exact opposite!

 

regards,

 

Martin.

 

Edit to remove format smileys.

[R A Watson]

 

(for younger readers there are 80 chains in a mile - so, for example, a 40 chain radius is a half mile = 880 yard radius or about 35feet radius in 4mm scale)

 

An easier method to remember, a chain is 22 yards or the distance between the wickets on a cricket pitch.

 

Wally

[big jim]

a modern day scenario is hatton (between warwick and dorridge) that is on a cant with a crossover between the up and down for trains coming from stratford upon avon, but also on the down line there is a trailing point from the goods loop and a facing point for trains going to/from plat 3 (stratford platform/loop line)

 

the linespeed is 70 for sprinter/freight etc and 85 for HSTs, down from 100mph before the station, the crossover between the down and up is the normal 15MPH however the crossover from the down into platform 3 and the exit from the goods loop is only 10MPH due to the adverse cant between the running main and link to the loop line which when you look at it with the naked eye is a good few feet higher than the main

 

i'll try and get some pics next time i stop there

 

 

[Donw]

Probably the easiest way to do this is to use an open topped baseboard make the mainlines on a ply trackbase and fix that at a cant. The crossover can then be built effectively flat but at an angle ( if you can follow me) any sidings etc. should be on a separate base laid without cant this will then show canted main lines and a flat yard which will give the effect you want.

Don

[coachmann]

I go along with your idea of open framing and canted plywood base Don. Alternatively, cork is an ideal material when properly glued down with PVA, as it can be formed with a Surform and sanding block to provide any angle one requires.

[Alister_G]

Thank you all as always for your very helpful and informative replies.

 

As I am trying to be as faithful as possible to the prototype (despite the inevitable compromises forced when modelling) then I think I will introduce some cant on the individual lines, and then join them for the crossover by bringing the down line slightly higher, so that the crossover is all in one plane. I will be using the suggestion of Donw and others in that the individual track beds will be created from separate plywood structures, and then a cork underlay. If you look at the photos of the prototype the main running lines are noticeably higher than the goods loop, and this would be the best way to achieve that.

 

The exit into the goods yard loop is slightly tricky, as it is formed by a three-way point, so I will have to experiment with how I apply the gradient to get back to the base level, but we are talking only millimetres so I would hope that using Larry's suggestion of sanding down the cork underlay to give a secure base to support the track will be sufficient. I may introduce a short length of straight between the slip and the three-way in which to apply the gradient.

 

Thanks again for your input,

[Miss Prism]

Don't forget cant doesn't scale linearly.

[RBE]

Of course it does. If the outer rail was 3" higher on the full size railway it will be 1mm higher on the model.

 

Cav

[Miss Prism]

Of course it does. If the outer rail was 3" higher on the full size railway it will be 1mm higher on the model.

 

That 1mm cant will be 76 times the amount required to balance the scale speed on the scale radius - tan theta = v²/rg

 

Sorry, I'm not responsible for the laws of physics, but they still apply.

[Orinoco]

Of course it does. If the outer rail was 3" higher on the full size railway it will be 1mm higher on the model.

 

Cav

Only if the gauge you are measuring from is to scale.

[RBE]

I am well aware of the physics as a qualified engineer and scientist however, we are not adding cant for physics purposes on our models we are adding it to look like the real thing not act like the real thing. We also dont have trains of scale weight, working suspension or anything else hence my dismissal of your comment. No offence intended just that it wasnt relevant to what is trying to be achieved. The narrower gauge of OO is very relevant however.

 

Cav

 

 

[Miss Prism]

My apologies. Being from a planet where scaling weight and suspension are important, I had mistaken this thread to be connected with making models act like the real thing.

[meil]

My apologies. Being from a planet where scaling weight and suspension are important, I had mistaken this thread to be connected with making models act like the real thing.

 

Then you have a big problem because you can't scale time. So your simple linear scaling simply wont work.

[Miss Prism]

I'm not attempting to scale time. I'm just referring to an object moving along an inclined curve such that the centripetal force is in balance.

[Edwin_m]

But if we lived in a world (universe?) where these things scaled then most curves on models would only be useable at a snail's pace.

[sulzer27jd]

I tend to agree with Coachmann on this. What modellers are trying to do is capture the 'impression' of the railway and thus often means resorting to all sorts of visual tricks ranging from selective compression of all sorts of things to 'selective exaggeration' of others. For instance in 4mm scale a 3inch cant lift is only 1mm while a lift of, say, an inch is hardly noticeable - so visual trickery has to come into play while almost none of us can hope to even halfway accurately portray prototype running line curvature.

 

Edit typo

 

If we go way back up this page to the quote above I think we get back to the reality of trying to make things look pretty. On the planet I come from we live in houses that costs lots of money so have only limited space for toys. My accommodation allows for a 9.5' radius, which would be ideal if I wanted to model a dock line with tight curves. Rather inconveniently that doesn't float my boat and i want to run main line trains (reasonably fast) so I accept that my curves will look too tight. On a positive note, most photographs that we see tend to make curves look much tighter than they really are so the difference between what we are used to seeing - in print at least - and what we model is not as bad as it might be.

 

In relation to cant, I agree that any curve on my layout that is to scale will require building an extension to my attic that will end up in my neighbours house and they might not want a model railway in their attic (as unbelievable as that is). What I did was to accept the rather tight look and place a packer of approx 1mm under the outer rail. When I put down the track there was a bit more PVA used than on the flat, so effectively the ballast holds the trackwork in place. Mmm sounds familiar.

 

Most of the time the curves at each end of the main scenic panel are not the focus of attention, which again works in my favour but they can still produce an image like this;

 

 

The techy science stuff is actually quite interesting and having build real railways (a long time ago) it makes sense, but there comes a time when we need to put down the manuals and just stick things down and play trains.

 

Good luck with the layout

 

j

[RBE]

I believe that was the point I was making too. No need to appologise miss prism. Cant on models are there for effect rather than necessity. My hst set I had back when I was a little nipper would easily go around 18" radius curves at probably a scale 200mph and never give a hint of coming off without any cant whatsoever. We dont need it but for effect we want it.

 

Cav