[WaltN]

Goku, on 26 September 2016 - 04:21 AM, said:

But what is wrong with current values?
https://en.wikipedia...ki/Grade_(slope)#Nomenclature
I have 2,3 and 4 options:
2 and 3 -> Permile xy.z = percent x.yz
4 -> the second box.

There is a great deal of confusion in the literature about what "run" is. Some say it's the distance along the hypotenuse of the right triangle. (After all, that's the distance the train "runs.") Others say "run" is the base of the right triangle. The sine of the angle will be rise/run in the first case, and the tangent of the angle will be rise/run in the second case.

Now, for typical railroad grades, the angle is small, and the sine, tangent, and angle (in radians) are approximately equal. But for those of us who model inclined plane railroads, the distinction is important.

So, without making a survey of how many in Case 1 and how many in Case 2, the important thing is for you to carefully and precisely define what "run" means for your editor. (All it takes is basic trigonometry to shuttle back and forth.)

By the way, the recent Windows 10 "Anniversary Update" (1607) put the kuybosh on my use of the MSTS RE. I'll be watching this project with a great deal of interest.

[Goku]

In my editor "run" is track length. It is hypotenuse, not triangle base.

[WaltN]

Goku, on 26 September 2016 - 05:53 AM, said:

In my editor "run" is track length. It is hypotenuse, not triangle base.

That's the way I think of it. Hooray!

Also, although the Wikipedia article initially says run is the horizontal distance, later it says, "Railway gradients are usually expressed in terms of the rise in relation to the distance along the track as a practical measure." The track follows the slope.

[Genma Saotome]

I think the key question is given a handful of examples from KUJU's RE, what are the equivalent values in this program?

[Goku]

WaltN, on 26 September 2016 - 07:32 AM, said:

Also, although the Wikipedia article initially says run is the horizontal distance, later it says, "Railway gradients are usually expressed in terms of the rise in relation to the distance along the track as a practical measure." The track follows the slope.

The difference is 0.03% for 2.5 percent gradient (25 permile is max for heavy rail in my country), so IMO thinking about it is not important. For 5 percent (50 permile - common for trams) it is 3% difference.

[Gman347]

I do not claim any expertise in railroad terminology but in the architectural profession run is considered to be the horizontal dimension and rise the change in height. The percent of gradient is then taken as the rise divided by the run. For a route builder the required vertical rise of the grade is a known as is the horizontal distance between the two points over which the rise occurs. These are the two meaningful dimensions and rise divided by horizontal dimension provides the percent of gradient. Alternatively, the horizontal distance needed to maintain a predetermined gradient is simply the rise divided by that gradient. The hypotenuse is a not really relevant unknown and calculating it is an additional unnecessary step. That being said it is true that for the relatively shallow gradient to which railroad right of way must be maintained the error in any case will be very small.

Just my two cents.

[Goku]

Gman347, on 26 September 2016 - 09:52 AM, said:

The hypotenuse is a not really relevant unknown and calculating it is an additional unnecessary step.

It is not additional step.
How it works now:

// set vector 0, 0, 1000:
float vect[3] = { 0, 0, 1000 }; 
// rotate it using track quaternion:
Vec3::transformQuat(vect, vect, track->qDirection);
// vect[1] is now permile value ( hypotenuse / rise ). 
// now calculate ratio:
oneInXm = 1000.0/vect[1];

[Gman347]

Please don't misunderstand my comment. I am not questioning how the program works, nor was I being critical. I was just making the observation that as a route builder the two things I would likely know are the distance between two points and the rise that is needed between those points and I would calculate the gradient as the ratio of those two numbers. I guess I am just remembering working with those variables in my model railroad days. Back then a maximum gradient would have been established and the total horizontal length of track needed was calculated based upon the desired rise. I would not need to know the hypotenuse of the triangle and it would have been an extra step for me to calculate it. I has always been my understanding that the locating engineers working for real railroads, planning a route in an area with changing terrain, basically dealt with the situation in similar fashion.

[Jovet]

Goku, on 26 September 2016 - 05:53 AM, said:

In my editor "run" is track length. It is hypotenuse, not triangle base.

I want to add that since MSTS uses fixed track pieces, this is a natural happening. In an ideal world the "run" would be the triangle base. But since doing that correctly would require numerous odd-length pieces of track, it isn't practical.

[Lindsayts]

Gman347, on 26 September 2016 - 09:52 AM, said:

I do not claim any expertise in railroad terminology but in the architectural profession run is considered to be the horizontal dimension and rise the change in height. The percent of gradient is then taken as the rise divided by the run. For a route builder the required vertical rise of the grade is a known as is the horizontal distance between the two points over which the rise occurs. These are the two meaningful dimensions and rise divided by horizontal dimension provides the percent of gradient. Alternatively, the horizontal distance needed to maintain a predetermined gradient is simply the rise divided by that gradient. The hypotenuse is a not really relevant unknown and calculating it is an additional unnecessary step. That being said it is true that for the relatively shallow gradient to which railroad right of way must be maintained the error in any case will be very small.

Just my two cents.

When railways were laid out in the pre computor days, surely the horizontal distance would NOT have been known. The line was laid out with a team of surveyors, pegging out the line using chains and a set of theodelites. For important projects particularly over difficult terrain a number teams were used and the results compared to get greater accuracy. Due to this method of laying out the line only the distance the line actual covered would have been known, ie in hilly terrain horizontal distance would have had to be calcaulated. As Goku has already mentioned the difference between the true track distance and the actual horizontal distance is not really significant. Note : if you set up a route using markers taken from Google earth as long as one can set the grades correct th actual track distance should not really be an issue.

The error caused by the simplistic world model used by MSTS in most cases would produce an error greater then this error.

Incedently I have measured on Google earth the actual horizontal distance (Note 1) of the line Between Melbourne in Victoria Australia to Wagga in New South Wales, The rail distance i something around 450 kilmetres the measured distance varied by less than 300 metres, this corresponds to an error of around 0.06% an amount that would be insignificant in a route that long.

Note 1: The length was measrured of all straights between the points when a curve ended and the next curve started, the length over curves was measured in 100metre or so sections.

Lindsay

[Gman347]

Quote

When railways were laid out in the pre computor days, surely the horizontal distance would NOT have been known. The line was laid out with a team of surveyors, pegging out the line using chains and a set of theodelites. For important projects particularly over difficult terrain a number teams were used and the results compared to get greater accuracy. Due to this method of laying out the line only the distance the line actual covered would have been known, ie in hilly terrain horizontal distance would have had to be calculated.

I assumed that the process would have started with a map, knowing horizontal distances. Interesting stuff. Considering how track might snake around in hilly terrain that explanation certainly makes perfect sense.

Sorry, I didn't ,mean to highjack this thread.

[Lindsayts]

Gman347, on 27 September 2016 - 04:22 AM, said:

I assumed that the process would have started with a map, knowing horizontal distances. Interesting stuff. Considering how track might snake around in hilly terrain that explanation certainly makes perfect sense.

Sorry, I didn't ,mean to highjack this thread.

A couple of points.............

I do not believe the thread was hijacked, the question of the distance as specfied on a railway is definitely something one needs to ask when doing a route, and in fact if the distance was a true horizontal distance or the actual track distance is something a had wondered about. How it was done was expalined to me by an old engineer/surveyor.

On your first point, while the majority of Britain was mapped by at least the 18th century, in places like the America's and Australia, the railways generally proceeded mapping so no distances where known. In hilly terrain often a number of paths were surveyed, all paths being pegged out. If the area has been undisturbed (say a native forest that has not been logged) on a very carefull search one can sometimes still find these pegs for the alternate routes, we (the group I work with) have found some of these.

Imagine what it would have been like surveying a line in mountainous terrain, photographs of these teams in Australia show they were around 20 people involved in each team, everything including any calculations had to be done by hand, even high precision slide rules not being generally availible till around the end of the 19th century.

LIndsay

[WaltN]

Lindsayts, on 27 September 2016 - 01:49 AM, said:

The line was laid out with a team of surveyors, pegging out the line using chains and a set of theodelites.

There are two things wrong with that statement: (1) A theodolite is used for measuring angles (vertical and horizontal planes) not distances, and (2) it is the chain (singular) that is used to measure distance -- horizontal distance.

As a pointed out on Monday, if you do a web search on "survey +grade" you'll find (1) run is measured along the horizontal, and (2) run is measured along the hypotenuse (the path the wheels roll). (The Wikipedia article "Grade (slope)" chooses Interpretation 1 but goes on to say, in railroading, it's Interpretation 2.) Then I said that for typical railroad grades, it's "macht nichts." Goku substantiated that.

I'm going to wax eloquent on this subject, but I need to establish my credibility.

• In 1957, I had a summer job working with a Pennsylvania Highway Department Surveying team. Among other things, we surveyed Interstate 80 in the Stroudsburg area. Well before the end of the summer, I was relieving the transitman and "running the gun."
  
• In the 1970s as an IBM engineer, I augmented an APL implementation of COGO with graphics output.
  
• With Open Rails, I proposed a "route designer" as a tool to accompany a route editor. The proposal went over like the proverbial "lead balloon."

I want to talk about three areas. The first two relate to surveying practice in highway and railway design and surveying. The third relates to an MSTS dirty little secret.

Route Design
Route design largely takes place in an office -- yesteryear and this year. Although CAD has supplanted the drafting table to some degree, it is 2D paper that will be used in the field. Most used are drawings in the horizontal plane. There, the route path is partitioned by stations. In the US, there is a whole-numbered station point every 100 feet. Also, there are construction points: start of curve, end of curve, intersection points, etc. that do not necessarily fall at 100-foot intervals.

The other type of drawing is the vertical profile of the route. Vertical profiles are plots of elevations (heights) of the route at distances along the route path. Given the relatively easy curves associated with highways and railways, the straight-line segments between segments are a reasonable approximation of the route path.

Field Work
Field work, done by surveying teams, is mostly stationing -- erecting stakes with pins at stations -- and leveling -- obtaining an elevation value to enable approximating the terrain.

Tools for stationing are the transit and chain. In 1957, a transit amounted to a small telescope on the top of a tripod. Also, there was a plumb bob centered beneath the scope, such that the scope could be located directly over the center of a pin or nail in a wooden stake marking a station. (As I recall, the head of the nail was 2 or 3 mm.) The base of the scope was levelable, and leveling was an important part of moving the transit from one point to another.

The preponderance of the distance measurements made with the transit were made in the horizontal plane. I only saw one measurement made along a slope all summer long. Trigonometric functions were acquired from a book of tables, and only the Party Chief carried one. Since calculation (multiplication) was long-hand and by the Party Chief, he rarely called for such a measurement. Measurements in the horizontal plane were done with the chain (graduated steel tape marked in feet and hundredthsin the US). One end was held on a stake nail with a plumb bob used at the other end. Occasionally, bobs could be used on both ends, but accuracy suffered.

Preliminary Summary
When railroads (real or virtual) are built with sectional track or rail of specific length it is only natural to use the length as the run. To do otherwise would require extracting a square root.

An MSTS Dirty Little Secret
A bunch of years ago when I first joined the Open Rails Development Team, I discovered that MSTS track sections are rigid bodies. That's an important realization for curves. Elevate a curve section at one degree say, and the outbound end of the curve is no longer level; it is only level at the pivot end. Of course curved sections are generally pretty short (short meaning limited angular span). So the tip is very slight.

However, dynamic track is different; you can make an angular span up to 90 degrees. Worse, dynamic track can have five subsections (straight, curve, straight, curve, straight), and the whole business is treated as a rigid section. Now (this will blow your mind), perform an experiment: Layout a single dynamic track section with subsections 0, R<90, 0, R<90, 0. That is, the straight sections are of zero length, and the two curved section are of radius R (whatever you want to make it) and 90 degrees in span. (Both curves have the same specification.) Finally, elevate the section at, say, 3 degrees. Question: If the height at the start of the section is H, what is the elevation at the end of the section? Answer: H. How can that be? Answer: The section (all subsections) is a rigid body. So what is the problem? We are led to believe the grade is constant (3 degrees) all throughout the path. It isn't. (The path does not follow the helical curve we expected.)

Open Rails, with its objective of matching MSTS behavior does just that. But, any new route editor should consider, in the future, handling curved sections the RIGHT way in addition to the MSTS way. (It would take an option setting, and that is the domain of Open Rails. It is important that the traveler and the geometry have the same behavior.)

Please forgive the length of this post.

[longiron]

WaltN, on 28 September 2016 - 06:53 AM, said:

An MSTS Dirty Little Secret
A bunch of years ago when I first joined the Open Rails Development Team, I discovered that MSTS track sections are rigid bodies. That's an important realization for curves. Elevate a curve section at one degree say, and the outbound end of the curve is no longer level; it is only level at the pivot end. Of course curved sections are generally pretty short (short meaning limited angular span). So the tip is very slight.

this phenomena is readily apparent when you try to lay 3T curves on an rapidly rising (or declining) gradient. I've experienced it first hand in the route I'm building. Snap to the inside track of the three, and the train experiences a 'bump' on the middle and outside paths while in inner is smooth. Snap to the middle track and the train experiences 'bumps' on either of the outside paths. If you get really close to the track joint you can easily see the mismatch. Lay each of the 3T individually, and you have to compensate the gradient to get individual sections to seamlessly snap to the successive 3T section. Total PIA.

chris

[WaltN]

I guess I don't know what "3T" curves are. But, if you can SEE the discontinuity, then it's in the geometry. The phenomenon I was referring to caused trains to gradually leave the geometry vertically throughout the curve section because the traveler was following a path different from the geometry. When the section end was reached, the train would slam back in contact with the rails. Now, both the dynamic track curve and traveler use the same code.