Problem with reading curve radius

[legia]

I have a problem when reading curve radius in Surveyor, using option "Tracks", "Show curve radius".

I laid down section of a track with radius of about 1500 m, but when I move mouse pointer over the track I get as many as four different readings, e.g.: 1279, 1456, 1590, 1877. I move the pointer few pixels and the value changes significantly as the track would jump few hundred meters away.

I believe my method of plotting the curve is fairly accurate and should not allow for error margin of up to 32% as the readings suggest. It looks to me as the error margin calculated in the Surveyor increases proportionally to the value of the radius. In curves of 4000 m and higher the margin can be so great, I doubt the readings are true.

Is there any tool to measure radius accurately?

 

[bl4882]

As far as anyone knows it is impossible to lay a curve of an even radius in Trainz. It is just a fact of life. You don't say if you are using a curve template. If you are not, you might search the DLS for one of the several that are up there. These can help to minimize the problem. If you can figure out a way to eliminate the problem, let everyone know, and you will probably receive a medal the size of a frying pan. :hehe:

You might also take a minute to register whatever version of Trainz you are using.

Bernie

 

[cascaderailroad]

As far as I know the measuring tool is only usable, where the numeral jumps from 0.00 , to the curve numeral (a few pixels away).

I use FT Track as a curve guide, and slide my track spline points over top of the FT Track.

 

[RPearson]

The trainz spline mathemetically behaves very much like a cubic bezier spline. What the actual formation is Auran has never indicated. Anyhow this type of spline can't represent a circular arc exactly. It can approximate one fairly closely if the points are equally spaced along the arc you want to match and the correct end tangents are enforced with straight section of track. If the spacing is around 30 deg or less the variation in the radius shouldn't be more than a couple percent. Manually it can be diffcult to do. But with all the layouts done to date a fair number of them have managed to get good looking curves put down on the terrain - so it is possible but probably not too many that are circular. You could argue that the rr curves incorporate transitions at the ends that are usually a type of cubic spline. Which sounds good but you have to do a lot more work trying to get the trainz spline to match a circular arc with a cubic transition.

With this spline the curvature or radius tends to vary continuously along the length of the curve - sometimes a lot. I think the measuring tool is accurate.

Bob Pearson

 

[backyard]

Horsefeathers...

You can make a smooth transition curve in Surveyor.

For most curves no more than three spline points are needed. The adjustment of the curve is done by removing the gradient set on yellow spine points(Remove Gradient radio button), then slide them forward or backward until the curve is smooth. When you are satisfied with the geometry, use the Height tool(H) to set the elevation.

You can tell if too many spline points are present when a "kink" develops on either side of the spline point. Sometimes you don't need but two, and they will not necessarily need to be at the beginning of the radius...that usually makes a smooth transition curve. This benefits the "sway" or "tilt" features, that simulate superelevation.

Using the Straighten tool(B) can help if you run the tool all the way from one tangent to the other through the curve, but overusing that tool can cause kinks to develop also. If you straighten a tangent just before the radius, it may eliminate the transition into the curve.

View the curve straight overhead and also use the Mini-Map to view the entire curve whilst editing.

Here is a Conversion Guide for radius in meters and feet. Bump the thread if you will...please...:wave:

 

[legia]

Method used to plot an arc

Thanks for many interesting responses! Perhaps I should explain more about the method I use to lay down curved track in the beginning.

For the arcs of small radius and large angle, e.g. 400 m and less, 56 degrees or more I use sections of fixed straight track equal to the radius value and the arc center. This method is quite accurate and I don't worry about the readings I get.

For arcs of larger radii it is difficult to locate arc center (unless you attach several otherwise unnecessary boards), so I use the ruler to plot the circle tangent and fixed track to plot the angle. The tangent is determined by radius and angle. Angle is chosen arbitrary to make sure that the spline is within certain distance from the point where the two tangent sections cross each other (denoted as 'f' in most textbooks). I found 3 meters to be a good value in most cases. It assures smooth motion of the train and reduces time spent on laying extra splines. So, for instance, for arc of 1500 m, calculated angle between splines is 7 degrees and tangent 91.7 m. For arc of 2000 m the values are 6 degrees and 104.8 m respectively. If you are really pedantic you can go for f=2m or even less, but then you will need to live really long life to have enough time to make all those extra splines.

The second method is fairly accurate. Unfortunately the ruler isn't a very good tool for this job. It does not indicate precisely the points it measures distance between and it has accuracy of only +/- 0.5 m. The good thing about this method is that I can make almost any track layout easily and quickly. This includes turnouts which really are combination of a straight track and arc track or two, or three arcs.

I do not use arc or turnout templates, because they are never available in sufficient variety to implement even simplest routes.

You might also take a minute to register whatever version of Trainz you are using.
Bernie

I use TRS2004, TRS2006, TC, TS2009, but I can't indicate this in my signature, because Auran is unable to fix some problem in my accounts.

 

[backyard]

Predictability tells me when I am wrong...

This arrangement of a spur line south of Lakeland, Florida flashed me as a good example of what Surveyor does...work the same way every time. That predictability is why I like Surveyor.

Shown, is a correctly placed turnout. Not too short, with a spline point just before the junction. The two diverging lines have spline points that overlay on the edges. These four points along with the Straighten Spline tool are very flexible for setting the geometry in both directions.

Only two spline points were used to set the radius of the curves. The Mini map should be set to show the entire curve or sets of curves to the next nearly tangent track...it is not uncommon to see several radii in the same curve, however the increase or decrease of radius should not be severe at normal track speed.

These small-industry locations have unloading facilities parallel to the track but separated by a fence or drainage.



Here is with the Topo overlay for a 'location guide'...

 

[legia]

In arc construction spline spacing is critical

Your example with arrangement of a spur line south of Lakeland, Florida is very instructive and I am sure it looks nice in the simulator. However, it is not clear to me if it addresses following problems:

1/ Trains moving in zigzag motion.
I found that if adjacent splines are too far away or radius is too small, trains do not run smooth. Many maps that are built in TRS2006 suffer from this problem. I don't know exactly what is the reason, but I believe it has something to do with plotting train direction at any given moment using coordinates of the nearby splines. If the splines are far away the result of calculation of the train movement may be not very accurate. If radius is large this problem does not occur, perhaps because curve with large radius is close enough to a straight line. Using trial-and-error method and simple geometric formula I found maximum distance S between splines:
R=200m, S=69.8m
R=400m, S=97.7m
R=800m, S=125.7m
R=2000m, S=209.4m
R=4000m, S=279.3m

2/ On prototypical railway it is recommended to separate two reversed arcs with a section of a straight track. The length of the track depends on the arc radius or train maximum speed. For simplicity in my routes I use length of 100 m. In your example I believe there are no arcs, but two parabolas. If you use tool "Get track radius" you should get high values in the middle section, but low values in the parts where curve joins the straight track. The low values determines train maximum speed for this line, therefore it is better to make curve more balanced, where you get similar values along entire length of the curve. To do this using trial-and-error placement of splines is difficult and time consuming.

Sometimes you need to widen the distance between two tracks to allow for and island-platform. How would you do this by hand? Using calculation this task is quite simple: If, for instance, the distance between the tracks is 4m and platform width is 8.3m (as most island platforms in TC) you need to widen track for 5.3m. You can do this using straight section 78.3m long, at angle of 1.4 degree and connected with two arches of 4000m radius. The total length of the s-track would be 273.7m so you know in advance if you won't ruin any buildings or collide with roads nearby. If you not happy with this result, you can quickly recalculate it. Doing the whole thing by manual adjustments of splines takes more time and does not produce very accurate results in my opinion.

 

[backyard]

...track speed...

Thanks Legia for your reply.

The curve shown is following a prototype curve on a 3-mile industrial branch line. The speed limit is 10-mph/16-kph. There is no superelevation.

I am following Topo maps, so the track is only laid on a line, tangents are straightened using the Ruler, relative position to another parallel tangent is set using the 5m or 10m grid.

I do place spline points at specific elevation points. Using the Straighten tool down the line on each section between splines all the way from the last tangent to the next tangent track through the curve works to form a smooth radius curve where needed.

There is a speed limit guide several posts down the page from the Curve Radius Guide for track speeds.

Whatever works to lay smooth operations works for me.

 

[cascaderailroad]

I made circles of FT Tracks (50m - 600m R), and attached them at every 5th spline point to Track Alignment Guide 50 Track, and connected those with Track Alignment Guide 10 Track.

Now I can drag a complete selection of FT Track radius circles around my route, and place it over a curve area, then lay real track, and slide the spline points over top of the FT Track spline points.

In also choose to to temporarily dis-join any track circle, and lay it over a curve area.