Curve offsets, and 'bend' transformations.

A long time ago, on the other forum, someone posted a set of equations for working out curve radius for various offset values.

Could someone here please consider reposting the formulae concerned?

I am also looking for information on how to generate 'tunnel' sections using a formulae (namely that of a torus), however I've got
stuck on how to adapt the torus formulae to allow generation of 'tunnel' sections with a vertical curve as well as a horizontal.

It would be much appreciated if someone could explain what for any given pair of radius (R,T) defined in the X-Z plane, and Y-Z plane respectively, the generalised 'bend' transformation for a given point {x,y,z} in a rectangular object would be.

Try this Alex,
http://www.virtualrailroader.com/diverge-calc.html

Theres a handy offsets table to download, I have one printed A5 double sided and laminated, essential part of dev toolkit.

(Maybe Derryck will link this in knowledge base??)

Yes, maybe once I get up and running again. I am having a break at the moment, I am considering further steps. More details via pm, if you want.

graymac wrote:Try this Alex,
http://www.virtualrailroader.com/diverge-calc.html

Theres a handy offsets table to download, I have one printed A5 double sided and laminated, essential part of dev toolkit.

(Maybe Derryck will link this in knowledge base??)

That whilst helping to work out 'centreline' divergences' doesn't help me at the moment ,as I can't seem to generalise to a consistent formulae so I can work out where the left and right hand rails should go. (so I can use different gauges and so on)

There are at least two situations to consider ,
i) where the 'curve' is a freeobj., and so the 'curve' is divering with respect to z. To get the Left and right hand rail positions isn't
as simple as just translating the distances given in the formulae above IIRC.

ii) Where the 'curve' is a rail(x) type.

Can someone here come up with a diagram and formulae so I can work out where to put the left and right 'rails' respectivly
(which is essentialy what Switch does for standard gauge trackwork) ?

Once that formulae exists, it can be used in conjunction with other formulae greymac and others are aware to workout trackwork in an unprecdented level of detail

In the meantime if anyone can understand these : http://www.norgrove.me.uk/permanent_way_notes.htm conversion of the formulae concerned into a 'Switch' program replacement would be much appreciated.

Try the Excell file 'kurveberechnung' from this site. /download/programs
You might find it useful, as the formula is there...

Derryck wrote:Try the Excell file 'kurveberechnung' from this site. /download/programs
You might find it useful, as the formula is there...

It is? I'm sorry to sound like a noob, but I'm not seeing how to get the Left and Right rails from that spreadsheet...

Diagram here:

http://upload.wikimedia.org/wikipedia/en/e/ea/Track_curve.svg

It's the subscripted Z values , as well as a general formulae for X_left, X_Right and X_c that is sought.