Reverse Curves with tangential piece (can we remove tangential piece)

We are connecting two 75' radius vertical reverse curves. The software is adding a 1' tangential piece between the reverse curves. Can we get rid of this tangential piece and have the two reverse curves connect directly. Where do we modify this?

We are just starting out so sorry if this is obvious.

Thanks,

Doug Kuhn, P.E.

What version are you using?

You can't make a reverse vertical curve. I just tried it in 2009 and you still can't. What I did was change the curve length from 400' to 399.99999'. That worked in 2009 but I never tried it in earlier versions.

It is the same on earlier versions. Vertical curves work from tangent to tangent, and not to or from another vertical curve. All you can really do is trick it by adjusting the parameters very slightly so that there is a very short segment between the curves.

Measure the distance from PVI to PVI that you want the reverse curve. Place a curve at the first PVI using a length of the total distance between the two PVIs. Draw a circle at the begining, end, and on the curve. Erase it. Repeat for the second PVI. Erase the tangent between the two PVIs. Adjust the first tangent to end at the start of the first curve. Add a fixed curve, three points snapping to the centers of the three circles. Repeat for the second curve. Adjust the last tangent to start at the end of the second curve. Walla.

http://i28.tinypic.com/2lavq4w.png

That's some impressive machinations, Brian.

Of course, I think it would be a lot easier if we could just do like we did in Land Desktop, for example put in two 100' vertical curves with PVIs at 4+00 and 5+00, and not have C3D throw a fit over the fact that both curves have endpoints at station 4+50. That convoluted process you outlined is the sort of thing that just reinforces the impression that C3D is difficult to use.

Let’s not mention that if you try to think outside the C3D’s parts catalog, you’re in for some real headaches.

Can someone who understands this rewrite it in an expanded form. The directions are not very clear to me.

Brian's explanation is a bit complicated but it sounds like it would work. As I mentioned above another option is to just shorten one curve length by 0.0001.