Rails, Who Curved Them Question

[Andy873]

An interesting but ambiguous phrase was used by the L&YR in their standard specifications for would be contractors intending in bidding for contracts to build lines for that railway company in the 1870s...

After it explains that the straight rails (21 feet in length) must be laid exactly that, straight and free from twists and slight bends, it states the following (exact wording and punctuation):

"and in the curves must be bent and worked to the arc of the curve."

What does this mean exactly?
The contractor has to take straight rails and bend them?
The company will supply curved rails but you have to ensure the curve is the exact arc required?
And in any case, how on site would you bend the rails to the exact arc?

The drawings as mentioned for any line by the L&Y would always start with No. 1 Plan and No.2 Section. Now as a contractor you need to know what arc of a curve is required and where, so I presume one of these two drawings (perhaps numbered 2a, 2b etc) would have an exact track diagram of the new route?

It must have been far more time consuming to lay curved rails, not forgetting of course you have to lay them in pairs with the inside rail having a tighter curve than the outside one but still maintaining that 4 feet, 8 and a half inches, and it must be even harder when a line is double track, i.e. four curved rails.

 

[D Williams]

The curves would be set out and pegged by the surveyors using standard survey techniques, which is a subject in itself. . All railway routes include lots of curves. All rails are supplied from the mill as straight lengths. Rail is quite flexible and setting the rails to the required line is easily achieved by ( in those days) barring into position with crowbars. Have a look at Youtube to see how it's done now. The inside rail ( the low rail in pway speak) will be shorter than the outside ( high) rail proportional to the radius of the curve. In order to keep the rail joints in line ( in the USA they didn't bother) some, but not all, of the low rails would be shorter. The GWR practice was one in six. These were cut to standard lengths to avoid lots of cutting when replacements were needed. Again the number of "short" rails depends on the length of the curve. I'm referring to standard gauge plain track here. I understand that some narrow gauge rails have to be formed to the required curve with mechanical benders. Points and crossings were produced to order by companies , of which there were many in the railway building boom, that specialised in this line of work and delivered to site in component parts for assembly. Sleepers were / are pre-drilled for the chairs or baseplates so achieving the correct gauge is not a problem. The larger railway companies of course usually had their own track works. In the olden days laying iron rails on stone blocks must have been another matter!

No doubt a pway engineer will give a more technical explanation.

 

[Taunton]

The company will supply curved rails but you have to ensure the curve is the exact arc required?

This sounds like the old Hornby-Dublo catalogue, where the packaged sets came with "12 curved and 2 straight rails".

 

[PG]

I'd like to thank both @Andy873 for posing the question which had never previously occurred to me and @D Williams for an answer that even a layman, i.e. me, can understand.

All quite fascinating now that I think about it!

 

[edwin_m]

Modern track doesn't need curving, except on the tightest curves such as on tramways. This is seen by the way a long continuous rail can be transported on a train of flat wagons, bending to suit the track it is passing over.

I guess with joined track the ends might need to be bent to prevent kinks.

 

[Ploughman]

In my days as a Tech with Track Renewals, I had one job that required the use of precurved rails.
That was at Crimple Curve between Leeds and Harrogate.
The curve is well known to be tight and below 180m radius thus also requiring check rails to be installed as well.
Both Up and Down lines of the entire curve were to be renewed over successive weekends approx 12 lengths each week.
The track was built up in a yard by a supplier of pointwork they did all the curving to specified radii.
The first train load of panels were then loaded to a train which had to run under Out of Gauge instructions due to the overhang of the panels by a set route from Doncaster to York then Harrogate and finally Crimple. The UP line was done first with panels laid in by Twin Jib Tracklayer.
The following week the Train had to arrive on site from Leeds with the overhang to the opposite side.

 

[billh]

Standard gauge bull head rail can be bent by means of a "Jim Crow" machine. This is a heavy iron horseshoe shape with a large acme thread screw in the middle, there are two claws about 30 ins apart and these grip the rail while the centre pusher is screwed down using a large bar through holes in the top of screw thread. Bending heavy rail is very hard work and skill is required to work out just how much bend to put into each "press" . the machine is then moved along the rail and process repeated. It can also be used to straighten bent rails. There is a modern equivalent that uses hydraulic pressure.

 

[DelW]

The drawings as mentioned for any line by the L&Y would always start with No. 1 Plan and No.2 Section. Now as a contractor you need to know what arc of a curve is required and where, so I presume one of these two drawings (perhaps numbered 2a, 2b etc) would have an exact track diagram of the new route?

When the contract drawings were being produced, the draughtsmen would have had access to a box of curve templates of many different radii, which would have been used in drawing the track plans. In my time (up to and including 1990s) these were plastic templates, about 300 - 500mm long, in radii varying from about 300mm to 5000mm. The real-world radius each represented would of course depend on the scale of the drawing. Earlier ones would have been in imperial measurements.

They were kept in a purpose made hardwood box, and were always called "railway curves" although ours were only ever used for drawing road alignments.

By the end of the 1990s they'd been superseded by CAD plotting. At the time of an office move in 2006, I was tempted to "acquire" our last set, but on opening the box, probably for the first time in several years, I found the plastic had deteriorated and in the end they went into a skip.

 

[Andy873]

The curves would be set out and pegged by the surveyors using standard survey techniques, which is a subject in itself. . All railway routes include lots of curves. All rails are supplied from the mill as straight lengths. Rail is quite flexible and setting the rails to the required line is easily achieved by ( in those days) barring into position with crowbars. Have a look at Youtube to see how it's done now. The inside rail ( the low rail in pway speak) will be shorter than the outside ( high) rail proportional to the radius of the curve. In order to keep the rail joints in line ( in the USA they didn't bother) some, but not all, of the low rails would be shorter. The GWR practice was one in six. These were cut to standard lengths to avoid lots of cutting when replacements were needed. Again the number of "short" rails depends on the length of the curve. I'm referring to standard gauge plain track here. I understand that some narrow gauge rails have to be formed to the required curve with mechanical benders. Points and crossings were produced to order by companies , of which there were many in the railway building boom, that specialised in this line of work and delivered to site in component parts for assembly. Sleepers were / are pre-drilled for the chairs or baseplates so achieving the correct gauge is not a problem. The larger railway companies of course usually had their own track works. In the olden days laying iron rails on stone blocks must have been another matter!

No doubt a pway engineer will give a more technical explanation.

That's a great explanation thanks, worded in a way even I can understand!

I'd like to thank both @Andy873 for posing the question which had never previously occurred to me and @D Williams for an answer that even a layman, i.e. me, can understand.

All quite fascinating now that I think about it!

No problem, it is quite fascinating when you indeed think about it.

In my days as a Tech with Track Renewals, I had one job that required the use of precurved rails.
That was at Crimple Curve between Leeds and Harrogate.
The curve is well known to be tight and below 180m radius thus also requiring check rails to be installed as well.

I've found some of the best sources of information are from people like yourself who has actually been there and done it, once more, another thank you. By the way, your link to the NYMR page isn't working!

Standard gauge bull head rail can be bent by means of a "Jim Crow" machine. This is a heavy iron horseshoe shape with a large acme thread screw in the middle, there are two claws about 30 ins apart and these grip the rail while the centre pusher is screwed down using a large bar through holes in the top of screw thread. Bending heavy rail is very hard work and skill is required to work out just how much bend to put into each "press" . the machine is then moved along the rail and process repeated. It can also be used to straighten bent rails. There is a modern equivalent that uses hydraulic pressure.

Some more valuable information here. With the exception of points then, it seems the L&Y simply supplied straight rails and it's up to you to do those curves to the specified arc.

When the contract drawings were being produced, the draughtsmen would have had access to a box of curve templates of many different radii, which would have been used in drawing the track plans. In my time (up to and including 1990s) these were plastic templates, about 300 - 500mm long, in radii varying from about 300mm to 5000mm. The real-world radius each represented would of course depend on the scale of the drawing. Earlier ones would have been in imperial measurements.

They were kept in a purpose made hardwood box, and were always called "railway curves" although ours were only ever used for drawing road alignments.

Again, fascinating. Using templates for the curves by the draughtsman once more shows how much standardisation was in place by railway companies, something I've mentioned before.

Thanks everyone, appreciated!

 

[Rescars]

This sounds like the old Hornby-Dublo catalogue, where the packaged sets came with "12 curved and 2 straight rails".

Decauville supplied sections of ready-made portable narrow-gauge track. Does anyone know if these sections were ever supplied with pre-formed curves?

 

[Taunton]

In order to keep the rail joints in line ( in the USA they didn't bother)

No doubt a pway engineer will give a more technical explanation.

Just to add a comment to this the USA rail joints are not parallel, even on straight track, they (approximately) alternate each side, or "staggered joints". USA standard rails are also shorter, at 39 feet, than UK, where 60 feet is a longtime standard, and when in the USA a periodic shorter low rail is required on a curve that was commonly done to 36 feet, again to keep standards in stock. Long welded rails are now as common there as here, but where there are traditional jointed rails the noise inside and outside a train is quite different.

There is regularly concern in the UK that this could cause significant rocking side to side, but it doesn't seem to happen unduly, even on poorly maintained track, and some USA branch lines have allowed track into much worse condition than anything in the UK.

 

[Rescars]

Just to add a comment to this the USA rail joints are not parallel, even on straight track, they (approximately) alternate each side, or "staggered joints". USA standard rails are also shorter, at 39 feet, than UK, where 60 feet is a longtime standard, and when in the USA a periodic shorter low rail is required on a curve that was commonly done to 36 feet, again to keep standards in stock. Long welded rails are now as common there as here, but where there are traditional jointed rails the noise inside and outside a train is quite different.

There is regularly concern in the UK that this could cause significant rocking side to side, but it doesn't seem to happen unduly, even on poorly maintained track, and some USA branch lines have allowed track into much worse condition than anything in the UK.

IIRC, Hamilton Ellis claimed that, compared with the UK's "clickety-clack", the sound made by US trains travelling over staggered joints sounded more like "chewing gum, chewing gum"!

 

[Ken H]

A transition curve is where the track goes from straight to the curve raduis over a distance. The cant would slowly increase too.
How were/are these set out?

 

[Taunton]

A transition curve is where the track goes from straight to the curve raduis over a distance. The cant would slowly increase too.
How were/are these set out?

There is a mathematical calculation, involving logarithms, which I believe was extensively developed by Gresley's LNER drawing office in the 1930s (someone must have been good at maths) when the Silver Jubilee high speed express started running, and the train got shocks when it ran onto curves without transition. It's a combination of progressively increasing both the horizontal curvature and the vertical cant.

I don't know how it is laid out on the ground for the track, or measured for maintenance, but am aware that modern asphalt laying machines for roads are capable of having this programmed into them so when driven along the road alignment they do the transition calculation required and lay the material out precisely curved and banked as specified. Roads are even more complex than railways as the impervious top surface has to incorporate the rainwater runoff calculations as well.

 

[Ken H]

There is a mathematical calculation, involving logarithms, which I believe was extensively developed by Gresley's LNER drawing office in the 1930s (someone must have been good at maths) when the Silver Jubilee high speed express started running, and the train got shocks when it ran onto curves without transition. It's a combination of progressively increasing both the horizontal curvature and the vertical cant.

I don't know how it is laid out on the ground for the track, or measured for maintenance, but am aware that modern asphalt laying machines for roads are capable of having this programmed into them so when driven along the road alignment they do the transition calculation required and lay the material out precisely curved and banked as specified. Roads are even more complex than railways as the impervious top surface has to incorporate the rainwater runoff calculations as well.

Is a transition curve a parabola then?

 

[edwin_m]

Is a transition curve a parabola then?

I believe it's a clothoid. I'm not sure what one of these is but web search will probably bring something up.

 

[DelW]

I believe it's a clothoid. I'm not sure what one of these is but web search will probably bring something up.

In the days when transition curves were hand-calculated or looked up from tables, I think spirals as well as clothoids were used. Alignments have been calculated on software for at least 40 years, so other curves may be used by those systems.

I don't know how it is laid out on the ground for the track ...

The traditional method was to use a theodolite set over the tangent point and turned through a series of deflection angles, and a steel tape to measure the chord length from one peg to the next. For a circular curve each deflection angle was the same (for a given chord length*), while for transitions each deflection angle changed with distance along the curve.

*deflection angle in minutes = 1718.9 x chord length ÷ radius, a formula imprinted in my brain for nearly 50 years.