Creating Centerlines in Carlson Survey

https://surveyorconnect.com/community/threads/cool-analytic-geometry-for-surveyors.331833/

I do something similar, but keep like to keep the radius something simple, like 175', 200', 225', etc. I don't think that it's that critical to fit it perfectly, 173.56', or something like that.

Lugeyser, post: 449586, member: 1249 wrote: I normally best fit my arcs to get a rough radius, then fillet my tangents, start with my best fit radius and trail and error making radius fatter or slimmer until it hits my shots well.

I use this when we are have a rural road and a right of way needs to be generated based on field conditions.

Takes a lot of time, seems there is probably a simpler solution. Appreciate your suggestions!

Draw>Polyline by Nearest Found
Edit>Polyline Utilities>Smooth Polylines

Navaran,

Is there a way to make the curves radial? This just seems to make chords? Thanks!

Lugeyser, post: 449586, member: 1249 wrote: I normally best fit my arcs to get a rough radius, then fillet my tangents, start with my best fit radius and trail and error making radius fatter or slimmer until it hits my shots well.

I use this when we are have a rural road and a right of way needs to be generated based on field conditions.

Takes a lot of time, seems there is probably a simpler solution. Appreciate your suggestions!

I don't use Carlson, but have found a method that works well for me:

1. Create tangents by creating a best fit line. Use all the points you think are on the tangent, plus a few more. When the best fit line is created, look at the residuals and throw out those at each end that are substantially off the best-fit line. Do this for all of the tangent sections.
2. Connect the tangents created by creating a curve that passes through a point in the curves. Examine the offsets from the points on this curve to the curve line. If things don't look good, create another curve passing through a different point on the curve. Examine the offsets until yo have what looks like the best fit. If you want a radius with an even foot create one with a radius close to that of the curve you liked the best .

I do kinda what Jim does, but the only difference is that I always use either degree of curve or tangent length, old habits die hard I guess, I don't really look at a radius length, it just follows along.

Lugeyser, post: 449872, member: 1249 wrote: Navaran,

Is there a way to make the curves radial? This just seems to make chords? Thanks!

Lugeyser, post: 449872, member: 1249 wrote: Navaran,

Is there a way to make the curves radial? This just seems to make chords? Thanks!

COGO>Best Fit

Lugeyser, post: 449872, member: 1249 wrote: Navaran,

Is there a way to make the curves radial? This just seems to make chords? Thanks!

COGO>Best Fit

You can go to the DRAW, ARC, 2 tangents and external. The only problem is you have to know the external distance and enter it manually. I couldn't find a way to just pick the distance , also you have to extend the tangents to get a P.I.

I also have been using fillet by trial and error to make nice tangent curves after locating points on curb returns, I just can never pick out the pc and pt when locating in the field and can see with the naked eye that the curves just don't look right.
It would be nice to pick the two tangents and one or several points on the curve with the fillet command.
I noticed that the Cogo> best fit needs to have a coordinate file open.

I know this is a Carlson thread, but I'd like to chime in with a note on civil 3d... If you are coding linework to be reduced through the survey database and there are tangents on both sides of the curve then you can use the 'OC' linework code on a single point on the curve and the PC and PT will be adjusted accordingly to form a curve that is tangent to the line in and the line out and pass through the "OC" point.

Most rural roads the curves are many compound curves because they were located by the grader operator in order to put in a road by eye and there being no surveyed stakes.
I am happy that they have kept it in between the margins and not cross boundaries as much as possible.
The margins are what I locate and use for road locations, they are more permanent.
I use many combinations of 3pt curves and many times a whole snagged stretch of road will come out as a bunch of straight sections when following the margins.
Get as close to a line or curve that with a simple line and curve because in a few years the motor grader will have moved it.

Randy Rain, post: 449980, member: 35 wrote: I know this is a Carlson thread, but I'd like to chime in with a note on civil 3d... If you are coding linework to be reduced through the survey database and there are tangents on both sides of the curve then you can use the 'OC' linework code on a single point on the curve and the PC and PT will be adjusted accordingly to form a curve that is tangent to the line in and the line out and pass through the "OC" point.

Yeah, that's a real nifty tool Randy, but it would be nice if there was a command that it could be done besides field to finish

billvhill, post: 450056, member: 8398 wrote: Yeah, that's a real nifty tool Randy, but it would be nice if there was a command that it could be done besides field to finish

Try this:

In Civil 3D on the home ribbon, draw category, under the curve pull down the curve through point option.

https://forums.autodesk.com/t5/autocad-civil-3d-forum/fillet-lines-through-point/td-p/3157260

RRain

billvhill, post: 450056, member: 8398 wrote: Yeah, that's a real nifty tool Randy, but it would be nice if there was a command that it could be done besides field to finish

And for vanilla autocad (with lisp functionality) here is a lisp routine that I found to fillet two lines through a given point. This is a simple lisp routine that doesn't have any error handling routines so be sure that your curve point falls within the red triangle shown.

RRain

In my experience rural highways did not have curved layouts, unless, they were done in ƒ??modern timesƒ?.

Research to find the layout, and, evidence thereof would be where I would place my effort. In colonial times we all know that surveyors used a compass and a chain. There were no curve calculations performed.

Research of a highway in a recent project led me to a thorough description from 1830 and one dating earlier than 1760. It took a long time to find the handwritten layout and more hours than I would like to read the description.

Modern surveyors are taking the easy way out by using the center of the current pavement to parse the outline of the boundary.

I donƒ??t know the age of the highway you are working on, but, around here ƒ??ruralƒ? means very old.

Randy Rain, post: 450072, member: 35 wrote: And for vanilla autocad (with lisp functionality) here is a lisp routine that I found to fillet two lines through a given point. This is a simple lisp routine that doesn't have any error handling routines so be sure that your curve point falls within the red triangle shown.

RRain

OOps I forgot to include the lisp code:

https://forums.autodesk.com/autodesk/attachments/autodesk/130/267303/1/ftp.lsp

Randy Rain, post: 450083, member: 35 wrote: OOps I forgot to include the lisp code:

https://forums.autodesk.com/autodesk/attachments/autodesk/130/267303/1/ftp.lsp

Thanks Randy, I'll try both ways

If I remember correctly the curve fitter routine in the CG dropdown works better than the Carlson approach for creating a nice clean centerline geometry.