Gradient rising expodentially

[bomberswarm2]

I'm getting a train out of a valley using a square spiral around the edges, I figured that a 2.2% gradient would be able to do it in 1 3/4 rotations, so I set the gradient, and applied it to the track. However after doing a section of the track I checked the gradient and the gradient rose with each subsequent track section, from 2.2% to 2.21% on the second section, right up to 10%+ at the end. Why did it not set each individual section of track at a 2.2% gradient. Why is this happening?

 

[oknotsen]

Did you apply the gradient to each section of track? So between each set of spline points? Including the last part?

Could you share a screenshot of the situation? Maybe we see something that you are overlooking.

 

[bomberswarm2]

I set the gradient to 2.2, then clicked on each section of track, just after each spline point. Upon checking again, the smaller the section of track, the larger the gradient. I first set the gradient on the longest straight section of track, and on the next straight section after the first corner, the gradient goes from 10+ on the corner pieces, back to 2.2. But since I'm setting a gradient, the track length should not matter, as a gradient of 2.2 will rise 2.2 metres over 100 metres of track, so a 1 metre piece of track should rise 2.2cm.

I'm using 2012 not TANE btw.

 

[BuilderBob]

I set the gradient to 2.2, then clicked on each section of track, just after each spline point. Upon checking again, the smaller the section of track, the larger the gradient. I first set the gradient on the longest straight section of track, and on the next straight section after the first corner, the gradient goes from 10+ on the corner pieces, back to 2.2. But since I'm setting a gradient, the track length should not matter, as a gradient of 2.2 will rise 2.2 metres over 100 metres of track, so a 1 metre piece of track should rise 2.2cm.

I'm using 2012 not TANE btw.

If you look at the heights of the spline points I think you will find that the average gradient between points is what you specified, but the spot gradient varies wildly along the length. This is on old problem that has never been addressed. It seems that the degree of curve affects the way that the gradient is applied to the new section. My theory is that this is because, when the gradient of the preceding section of track is calculated, the calculation is the projection along a line from the spline point at the start of that preceding section, rather than as a projection along the instantaneous direction of the track at the end spline point. For curved track, the straight-line distance between spline points is less that the actual track length, so the calculated angle at the end of the previous section (based on the straight-line distance) is greater than the actual track gradient at the end point, and the new section is attached using this exaggerated gradient, but the transition is smoothed. The combination of an exaggerated required initial gradient and the smoothing needed creates a gradient that shoots up immediately after a spline point. The gradient for the remainder of the section is then adjusted to try to get the average back to what the user nominated. The result is a roller coaster. For a more detailed discussion see:
https://forums.auran.com/trainz/showthread.php?90049-Gradient-Problem/page2&styleid=7
especially the last summary.

 

[Dinorius_Redundicus]

Exponentially, not expodentially. There, my day is complete.

 

[bomberswarm2]

If you look at the heights of the spline points I think you will find that the average gradient between points is what you specified, but the spot gradient varies wildly along the length.

A 10 metre segment of track rising 2 metres isn't a 2.2 grade.

 

[BuilderBob]

A 10 metre segment of track rising 2 metres isn't a 2.2 grade.

Your complaint was that when you checked the gradient it was much higher than you expected, and it kept increasing for each section. If that's what really happened then my explanation is appropriate. If you really meant to say that you checked the heights of the spline points and they turned out much higher than you expected, given the reported gradient for the track, then that is entirely different and you can ignore my comment.

 

[cascaderailroad]

When applying a gradient, you constantly have to go backwards, and recheck your previous gradients, as they may have changed. Lay 6 gradients, go back 6 and recheck that they have not moved ... Lay 6 more gradients, go back 6 and recheck that they have not moved ... etc ...

 

[bomberswarm2]

I ended up fixing it by going over the entire slope about 6 times.

 

[Vern]

Try doing it with a spiral tunnel, get's even more funky. Click just beyond each spline point (from the surface, as you can't dive under the terrain), everything looks good at 2% or whatever. Drive it and suddenly you're going up a 6 or 7% hill. I also had this happen to me when attempting to build the Tamalpais or Mount Lowe (can't remember which) route and it got so frustrating I just abandoned the project.

 

[JCitron]

If you use T:ANE once you set a grade, all track put down after that will be at the at grade you entered. This can lead to some rather unintended results which I found the hard way.

 

[blueodessey]

I agree with all that's said here on Grades issues of compliance, one of my largest issues except for Faulty Asset Stuff, is I have to go back and manually check my Grades 3-4 times as I build long sections in Mountain areas of my Route or even in some low valley's etc, I find after I've adjusted a Track ahead for Grade, some how, some way, the Grade gets out of consistency, and like you all said, I end up with 2-3% difference, usually to high! :'( which means after running my Test Train, I go back to Surveyor and recheck all the change points....

It's Crazy but the nature of the beast, I'm used to it, but never like having to redo what I already had set perfectly according to the notes on heights I jotted down, ya, I'm old school, I still use pencil and paper.

Like a good carpenter once told me, measure twice, and cut once! Still follows true here.........