Oxygen molucule - Muffin Tin radius too big?

[kanato]

Has anyone used exciting for calculations on O2, or any compound which contains pairs of O atoms separated by roughly the O2 bond length?  (LiO2, KO2, etc)  I ask because I am doing calculations on RbO2 and I get this error:

Error(checkmt): muffin-tin tins overlap for
     species    2 atom    1
and species    2 atom    2

Sum of muffin-tin radii :    2.900000000
Distance between atoms  :    2.542723119

The distance between my Oxygens (2.54 Bohr) is slightly larger than the bond length in O2 (2.29 Bohr).  I can "fix" it by turning autormt on, but is there any problem associated with this?  Are there other basis parameters I need to change if the muffin tin radius is made smaller automatically like this?  I am not an LAPW expert..

[ckoe]

Dear Kanato,

at least the convergence with the number of plane-waves should
be slower, e.g. you need more plane-waves for the same accuracy,
if the muffin-tin radius is reduced.

Please see also the discussion about the importance of rgkmax in
http://www.fys.kuleuven.ac.be/iks/nvsf/publications/DFT_and_LAPW.pdf
as mentioned by Andrei in another thread.

Best Regards

Christof