Band gap in metallic carbon nanotube

[perfetti]

Dear every one,
       I am calculating the bandstructure of cnt(5,5), and it should be very easy. However, I always got a small bandgap near Fermi level in the bandstructure.
       I used k point =30,  50, 100, and it doesn't change much. I keep using GGA, PBE,  since I've no idea how I should set up about other exchange-correlation function.
      My input script and the bandstructure are attatched. Any advice should be appreciated.  I just want to eliminate this gap, or at least understand it.
      Thank you!

[Anders Blom]

Increase number of points on the band structure to 100. With just 20 points you don't hit the precise point where the gap is zero.