[Wien] QTL-B in *.scf vs. *.scf2

[cbridges]

Dear Users,

I have recently managed to converge a calculation, but with warnings
about the QTL-B value. I am confused about the different values
reported in case.scf and case.scf2. The values were taken from the
last cycle in case.scf, and the bottom of case.scf2.

case.scf2: QTL-B VALUE .EQ.    9.67978   !!!!!!
case.scf :  QTL-B VALUE .EQ.    1.82283   !!!!!!

I presume that one is an average value (in case.scf?)? Could someone
please explain what the difference is? From the UG, it seems that the
~1.8 value is nothing to worry about, but the ~9.6 value is a problem.

I have looked for the presence of ghostbands, and I can't find any
evidence for their presence. More specifically, I could not find any
large Q(UE) (energy derivative part) in the case.output2 or case.help*
files.

My problem could relate to the fact that both lanthanum and hydride 
(H-) are present in the system (and a 3d TM). The RMT values are 2.25
(La) and 1.4 (H-), with RKmax=7 and lmax=10, and all ions APW+lo. I
realize that this gives a much larger effective RKmax for the
lanthanum, but I was hoping that as the nearest neighbor distance is
much larger in the case of the hydride than for the typical hydrogen
case specified in the UG, the above parameters would be suitable.

My question is simply, are the QTL-B values above large enough that I
should worry? If so, should I then decrease the RMT of La to ~1.8, or
follow some of the other suggestions from the mailing list such as
reducing RKmax or changing La to LAPW?

Thank you for any suggestions you may have.
Craig