Track spacing, curve radius and speeds.

A collection of tutorials, guides and FAQ's.
Forum rules
Please only create new topics if you are posting a new guide, tutorial or FAQ.

Feel free to post in any existing topic asking questions, seeking clarification etc.

Track spacing, curve radius and speeds.

Postby Nobkins » Sun Apr 20, 2014 1:03 pm

Just a link to an excellent post by AndiS giving lots of very useful info on track configuration.

Feel free to extend the info in this thread.
TrainSimDev.com YouTube Channel
Video tutorials and other resources
User avatar
Nobkins
Site Admin
 
Posts: 547
Images: 3164
Joined: Wed Feb 05, 2014 12:24 am
Has thanked: 325 times
Been thanked: 185 times

Re: Track spacing, curve radius and speeds.

Postby AndiS » Sun Apr 20, 2014 11:34 pm

Now that you started a thread on the topic, I'd better fill it a bit systematically. The other thread is about fixing welded flangeways. In that context, I quoted track distance as a precondition but there is a bit more to it.

I must admit that I can be a bit nerdy when talking about prototype details, and I never finished a real route. But someone said (about project management): "You need a plan so you know by how much you are off." So take the following in this sense. Also consider that much is just as much work with the right figures in mind as it is with the wrong ones, and some frustration does stem from unrealistic model details.


Track distance

UK measure track distance between the rails near each other (right rail of left track to left rail of right track). The following are the basic numbers, translating from ft between tracks to metres between track centres:
6 ft = 3.4 m between running track pairs
10 ft = 4.6 m minimum between sidings
12 ft = 5.2 m between track pairs with signal gantry between them

In curves this is extended by a bit because the canted vehicles need a bit extra room. For RW, just forget about it, but maybe make the 3.4 3.5 in general, if you like. High-speed routes have their own rules with track distance depending on the speed.

A few German figures to compare:
3.5 m in old installations
4.0 m in not so old installations (or upgrades from 3.5)
4.5 m between sidings and between newly laid (high-speed) track, non-withstanding high-speed extra margins.

In stations or at junctions, the space between the pair of running track is generally maintained, but other times widened (e.g. from 3.5 to 4.0 m for German prototype) because the space is needed for crossovers and a bit of infrastructure on the ground.

On former broad gauge routes, platforms are spaced far from each other where they were retained when gauge changed from 7 ft 1/4 in to 4 ft 8 1/2. Assuming they used 6 ft between tracks (which I am not 100% sure) and they kept the outer rails (those next the platform), this would mean extra space of 4 ft 7 1/2 in on top of the 6 ft track distance. For practical use in RW, just remember that 10 ft is also a good value for track spacing between platform tracks on former broad gauge lines - where the station was not altered in 20th century.

Track distance 3.14 is a historic tragedy within the RW community. Someone at Kuju overheared someone talk about 3.4 and when he defined the first track rules, he took 3.14 from his memory. This is my theory, not inside knowledge. 3.14 cannot be explained otherwise, the space between rails would be called five foot if 3.14 was the right figure. Adding 6 ft to the standard gauge (leaving out rail width) leads to 3.26. So just put 3.14 behind you and fix your track rules before you start the next route.


Minimum radius

The following is backed by document GCRT5021 from rgsonline.co.uk. They are my favourite source because they aim at combining historical installations with modern thinking, e.g., risk assessment.

Minimum figures are:
200 m where passenger traffic occurs
150 m otherwise
125 m in exceptional cases


The relation of radius and speed in switches

The curves in switches are not compensated by cant. (This has nothing to do with RW's inability to combine superelevation and switches.) From the permitted cant deficiency, you can calculate the maximum speed. Cutting out the math, these are my results:
150 m radius, 23 mph
200 m radius, 27 mph
250 m radius, 30 mph
300 m radius, 33 mph
400 m radius, 38 mph
500 m radius, 42 mph

Combining these figures with speed limits found in Sectional Appendices and other notes from the period in question gives you an estimate of the radius of the curved track in switches. E.g., crossover of facing points between relief line and main line 40 mph in places where lots of through traffic is expected, 20 mph elsewhere (only used to get goods trains out of the path of a passenger train).
In this example, we can theorise that the 40 mph crossover had a radius of 500 m while the 20 mph crossover's radius was 150 m, which confirms that it was only used by goods trains.

German figures (radius and speed)
190 m, 40 km/h (25 mph)
300 m, 50 km/h
500 m, 60 km/h (38 mph)


Angles of switches

The track beyond the diverging track of a switch forms an angle with the straight track (assuming that the other track is actually straight). This is usually given as a ratio. 1:7 is quite a wide angle, but it seems common in UK for freight-only trackwork. 1:9 is the German standard angle for yards and crossovers up to 50 km/h.

I found references to 1:5.5, too, but I don't know how widespread that is. And I am sure a crossover between track spaced at 3.4 m would need a radius of 100 m to reach this angle.

UK tradition does not categorise switches like the German (or US) one. Still, these figures can serve as a guideline.

For all you care in RW, the angle expressed as ratio equals the ratio of curve length and radius. 21.4 m of curve at a radius of 150 m turns the track by 1:7. This means that the straight following this curve will lead away from the mainline from which it is forked off at a rate of 1 m for every 7 running metre.

The fact that using an 1:9 angle, it would be 9 instead of 7 metres for 1 metre of sideways offset means that for any given offset, i.e., for any given track spacing, the smaller the angle of the switch, the longer the straight between curves forming the crossover. This is an important point regarding the guide rail situation in RW.

If you don't have a straight between curves, the situation is the same - smaller angle = bigger denominator in the ratio = more space for the guide rails, and more leeway in placing the splitting point for the split-and-reweld technique.

On the other hand, a crossover at 1:9 is longer than a crossover at 1:7. A raw estimate for the length of the former is 60 m, compared to 50 m for the latter.


Straights in reverse curves

The two curves of the switches in a crossover form reverse curves. RGS say that "A length of straight track not less than 3 m long shall be provided between the reverse curves if one of the curves has a radius of less than 160 m."

The idea is that the wagon rotated left should stop that movement before it gets rotated right.

You can do that in RW. I tried it out: Two curves of about 21.5 m length and some 150 to 160 m radius gives a bit under 4 m of straight between them in a crossover between track spaced by 3.4 m. The point is not to memorise these figures. The point is that can have 1:7 crossovers in-game, featuring the straight prescribed.


The dangerous attraction of the minimum radius

Laying any curve using the minimum radius is very tempting - on have one thing less to worry in this awful process of fumbling with the mouse reading multiple figures that change while you look. However, there are several pitfalls.

1) When you lay an S-curve (like a crossover) and the first curve's radius is at the minimum, the second curve, which will never have precisely the same radius, can only deviate in one direction - towards a bigger value. If you start at another value and aim the second curve at the same one, the unavoidable deviations will be on one of two sides (bigger or smaller radius). This halves the average deviation in practice.

2) When you lay multiple track, at the minimum radius, there may be issues with the easement calculation.

3) When you join two curves intersecting each other at an angle, you need a tighter curve to do it. If those curves' radii is near the minimum, you will not be able to fit the joining piece in.

Focussing on 1), I suggest that you define your track rule with a minimum radius of 10 less than what you intend to use in practice. Or you make your curves some 10, maybe 5 m wider. How much of a leeway you need depends on how precisely you are working.

The ultimate aim here is to create curves at a radius that is similar to what the snap-to-track tool gives you when you use it to complete the crossover (or similar pieces of track).

Of course, it is easy to form a crossover by a curve at minimum radius followed by whatever the snap tool suggests, be it twice that radius. However, I don't like the look of such crossovers and they are longer than required, or to tight for their length.


Single and double slips

This is a bit of a black art in RW. I feel that UK people could save some nerves by trying things "the German way" if needed. The idea is this: For any given minimum radius, there is a certain range of angle at which the slip can be formed. It is easy to form this angle in the switch in the parallel track, where the diagonal track starts. I.e., you decide on an angle, e.g., 1:9 and a radius for the first switch, e.g., 200 m. You compute the length of the curve in that switch as 200 / 9 = 22.2. You form the curved branch of the switch at about 200 m radius and about 22.2 m length, and follow it by a fairly long straight. Then, you form the double or single slip at the intersection with the neighbouring track. Last, you split off the straight a bit beyond the slip and replace it by the curve you want to have (keeping a short piece of straight track to be sure that you don't damage your work).

It is clear that UK practice is more elegant, drawing a seamless the curve from the switch over the parallel track into the diverging route. But I guess that people often err in their estimation of the angle of the slip in such cases and push the route editor into a position where it must fail.

The other option that you have is the minimum radius. By switching between track types (mainline, passenger, freight, yard), you should have different radii available, if the track rule is good. I would not say that the route editor creates the curved parts of the slip exactly at the minimum radius, but it certainly influences it. Remember to change the track type back to mainline if you used passenger or freight type, as you would lock out half the AI otherwise.

When something goes wrong in a slip, use Undo immediately. Keep lots of copies of your work as you work on slips.

When I need to delete a slip later, I always drag up the white circle to select all its parts. The four pieces of straight track and the one or two curves form a single entity internally (in the data structure). However, the route editor sometimes feels challenged to locate them all and remove them if you remove some of these parts, which is easily done my clicking on one and deleting that one.

If you moved a part by accident, use Undo immediately, or delete the slip and lay it anew.


Platform Distances

Researching the above stuff, I came across this related topic.

GC/RT5204 states that from the inner rail edge to the platform edge you got 730 mm, minimum.
Together with a gauge of 1435 mm, we arrive at 1.448 m from track centre to platform edge which I tend to round giving 1.5 since this is the precision the offset tool shows.

The offset to enter in the route editor is thus 1.5 plus half the width of the platform, if the platform origin is on the centre line of the platform loft. For the default platforms in RW, this results in the following offsets to enter when laying them using the offset tool.
Platform width 5 m - offset 4 m
Platform width 7.6 m - offset 5.3 m
Platform width 9 m - offset 6 m

Platform height measured from rail top is 915 mm (+0/-25mm), or 0.9 m in RW precision. After placing the platform using the offset tool, the platform surface will align with the rail tops (vertically). Add 0.9 to its Y value and you are spot on the standard.
Edit: Added warning about deleting slips.
AndiS
Top Link Driver!
 
Posts: 736
Joined: Wed Apr 09, 2014 5:48 pm
Has thanked: 268 times
Been thanked: 308 times

Afterthoughs on single slips

Postby AndiS » Mon Apr 21, 2014 11:24 am

The situation most hurting the UK rail fan is a single line diverging from a double line. In the prototype, the single line would join both tracks of the double line in elegant, continuous curves. In RW, we hit two limitations at once:

  1. For all I know, RW does not form single slips (or double slips) at crossing where one of the tracks is curved.
  2. Superelevation on the single line must start from a straight near the single slip. So the curve is broken by this required straight and even more by the start of the easement which looks just as straight.
I don't expect DTG to fix both issues any time soon, so we need a plan to live with it.

In general, I prefer to start by laying the diverging switch. This gives me control over the switch geometry. If the curved track take too much space, i.e., if it protrudes into the space where the slip will be formed by the game, you are shot. However, joining the diverging track with pre-existing parts of the single line (where it is constrained by other reasons) can be a nuisance, too.

In such a case, the following procedure works just as well:
  1. Design the approach of the converging single track in such a way that it hits the double track line at an angle between 1:7 and 1:12.
  2. To determine this, extend a straight from the converging track across the other tracks (and a fair bit beyond).
  3. Using the measuring tape, find the distance between the intersections of rails of the diagonal track and one of the two others. Dividing this number by 3 gives the angle, e.g., 36 m corresponds to 1:12, 21 m corresponds to 1:7.
  4. If your number is outside this range, modify the curve before the straight to adjust the angle. A difference of 1 in the denominator of the angle (e.g., from 1:10 to 1:11) roughly corresponds to a curve length of 1/100 of the radius in your approaching curve. Just to give you an idea of how much to add or prune there.
  5. 1:7 is more for freight-only track, 1:12 corresponds to 500 m switch radius and 40 mph in the junction, just to give you an idea. So you may want to fine-tune the angle if you barely hit these limits.
  6. If you are pleased with the angle, form the single slip at the nearer track. Remember to use a track type (mainline vs. yard) with a suitable minimum radius - 1:7 with more than 200 m minimum radius may not work. 1:12 with 150 m minimum radius will look stupid and deliver a rough ride on the diverging track.
  7. Form a diverging switch at the diagonal track just beyond the guiderail of the crossing of the single slip, aiming at the second track using snap-to-track.
  8. Delete the unneeded straight (the end of the diagonal) and weld where the red portal shows (which is the normal process of forming a converging switch).
AndiS
Top Link Driver!
 
Posts: 736
Joined: Wed Apr 09, 2014 5:48 pm
Has thanked: 268 times
Been thanked: 308 times


Return to Tutorials, Guides, FAQ

Who is online

Users browsing this forum: No registered users and 2 guests