Your calculations are correct. X=R(1-cosø) where X = the offset, R = the radius of the circle, and ø = portion of the circle the section the track represents. Plug in 22.5º and 1550mm and you get 118mm. With back to back identical curved sections the total offset is 2X or 236mm. I suspect the radius is actually "relaxing" when you set this up with real track.

Thank you, Phils2um and Purellum!

I too, thought that it might have been my cutting. That's why I also tested the not-cut 59076 and made sure that they really clicked well into position. I cannot put a paper in the joints. And I measured along the planks in the wooden floor.

I like the hypothesis of a "relaxing" track. I will see when I get enough of them to make a circle, how the ends meet. I assume the circle would be open and that I would have to push them together. That the radius is a little more than 1.55, kind of. Stay tuned. I will update here.

Still I am rather confused about why this is so.

Also, strange that the picture of the 59076 is different from what was in the box.

Øyvind

Ok, there's something called a cutting disc? Please explain.

I cut mine like I wrote in the blog note:

If I after the cutting place one on top of the other they are as equal as I am able to observe. But since I too was suspiscious about that, that's when I checked a non-cut 59076 and found a large discrepancy there, too.

For my half 59076 then 41 mm instead of the 60 mm theoretical is quite ok on my shelf. But if I really wanted 60 mm, I could, as you say, file it. But that would have looked strange, because now the edges are nice and smooth.